daily pastebin goal
86%
SHARE
TWEET

Untitled

a guest Feb 13th, 2018 56 Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
  1. $ python run_job.py -n 5 -g 60 -c 12 --simulator_procs 10 --use_sync --name breakout --short
  2. args.offline:  False
  3. ('bash command: ', 'srun -A luna -N 5 -n 5 -c 12 -t 55:00 distributed_tensorpack_mkl.sh 61216 30065 Breakout-v0 adam 1 "3nodes 12cores" "breakout_1517329773.87" 0.00015 128 60 0 None 256 10 1 uniform normal False . True 1 False /net/archive/groups/plggluna/intel_2/logs/ 1e-08 0.9 0.999 0 False False False False 120 False')
  4. srun: job 9511741 queued and waiting for resources
  5. srun: job 9511741 has been allocated resources
  6. SLURM_JOB_ID  9511741 ; SLURM_JOB_NAME  distributed_tensorpack_mkl.sh ; SLURM_JOB_NODELIST  p[0111-0112,0115,0574,0576] ; SLURMD_NODENAME  p0112 ; SLURM_JOB_NUM_NODES  5
  7. SLURM_JOB_ID  9511741 ; SLURM_JOB_NAME  distributed_tensorpack_mkl.sh ; SLURM_JOB_NODELIST  p[0111-0112,0115,0574,0576] ; SLURMD_NODENAME  p0111 ; SLURM_JOB_NUM_NODES  5
  8. SLURM_JOB_ID  9511741 ; SLURM_JOB_NAME  distributed_tensorpack_mkl.sh ; SLURM_JOB_NODELIST  p[0111-0112,0115,0574,0576] ; SLURMD_NODENAME  p0115 ; SLURM_JOB_NUM_NODES  5
  9. mkdir: cannot create directory ‘/net/archive/groups/plggluna/adam/experiments/breakout_1517329773.87’: File exists
  10. mkdir: cannot create directory ‘/net/archive/groups/plggluna/adam/experiments/breakout_1517329773.87’: File exists
  11. mkdir: cannot create directory ‘/net/archive/groups/plggluna/adam/experiments/breakout_1517329773.87/models/’: File exists
  12. mkdir: cannot create directory ‘/net/archive/groups/plggluna/adam/experiments/breakout_1517329773.87/models/’: File exists
  13. mkdir: cannot create directory ‘/net/archive/groups/plggluna/adam/experiments/breakout_1517329773.87/storage/’: File exists
  14. mkdir: cannot create directory ‘/net/archive/groups/plggluna/adam/experiments/breakout_1517329773.87/storage/’: File exists
  15. SLURM_JOB_ID  9511741 ; SLURM_JOB_NAME  distributed_tensorpack_mkl.sh ; SLURM_JOB_NODELIST  p[0111-0112,0115,0574,0576] ; SLURMD_NODENAME  p0574 ; SLURM_JOB_NUM_NODES  5
  16. SLURM_JOB_ID  9511741 ; SLURM_JOB_NAME  distributed_tensorpack_mkl.sh ; SLURM_JOB_NODELIST  p[0111-0112,0115,0574,0576] ; SLURMD_NODENAME  p0576 ; SLURM_JOB_NUM_NODES  5
  17.  plgrid/libs/qt/5.4.1 loaded.
  18.  plgrid/libs/qt/5.4.1 loaded.
  19.  plgrid/libs/qt/5.4.1 loaded.
  20.  plgrid/libs/mkl/11.3.1 loaded.
  21.  plgrid/libs/mkl/11.3.1 loaded.
  22.  plgrid/libs/mkl/11.3.1 loaded.
  23.  plgrid/tools/gcc/4.9.2 loaded.
  24.  plgrid/tools/gcc/4.9.2 loaded.
  25.  plgrid/tools/intel/15.0.2 loaded.
  26.  plgrid/tools/intel/15.0.2 loaded.
  27.  plgrid/tools/gcc/4.9.2 loaded.
  28.  plgrid/tools/intel/15.0.2 loaded.
  29.  plgrid/tools/tcltk/8.5.19-threads loaded.
  30.  plgrid/tools/tcltk/8.5.19-threads loaded.
  31.  plgrid/tools/python/2.7.13 loaded.
  32.  plgrid/tools/python/2.7.13 loaded.
  33.  plgrid/tools/tcltk/8.5.19-threads loaded.
  34.  plgrid/tools/python/2.7.13 loaded.
  35.  plgrid/libs/qt/5.4.1 loaded.
  36.  plgrid/libs/mkl/11.3.1 loaded.
  37.  plgrid/tools/gcc/4.9.2 loaded.
  38.  plgrid/tools/intel/15.0.2 loaded.
  39.  plgrid/tools/tcltk/8.5.19-threads loaded.
  40.  plgrid/tools/python/2.7.13 loaded.
  41.  plgrid/libs/qt/5.4.1 loaded.
  42.  plgrid/libs/mkl/11.3.1 loaded.
  43.  plgrid/tools/gcc/4.9.2 loaded.
  44.  plgrid/tools/intel/15.0.2 loaded.
  45.  plgrid/tools/tcltk/8.5.19-threads loaded.
  46.  plgrid/tools/python/2.7.13 loaded.
  47.  plgrid/libs/mkl/11.3.1 unloaded.
  48.  plgrid/libs/mkl/2017.0.0 loaded.
  49.  
  50. The following have been reloaded with a version change:
  51.   1) plgrid/libs/mkl/11.3.1 => plgrid/libs/mkl/2017.0.0
  52.  
  53.  plgrid/libs/mkl/11.3.1 unloaded.
  54.  plgrid/libs/mkl/2017.0.0 loaded.
  55.  
  56. The following have been reloaded with a version change:
  57.   1) plgrid/libs/mkl/11.3.1 => plgrid/libs/mkl/2017.0.0
  58.  
  59.  plgrid/libs/mkl/11.3.1 unloaded.
  60.  plgrid/libs/mkl/2017.0.0 loaded.
  61.  
  62. The following have been reloaded with a version change:
  63.   1) plgrid/libs/mkl/11.3.1 => plgrid/libs/mkl/2017.0.0
  64.  
  65.  tools/gcc/6.2.0 loaded.
  66.  tools/gcc/6.2.0 loaded.
  67.  tools/gcc/6.2.0 loaded.
  68.  plgrid/libs/mkl/11.3.1 unloaded.
  69.  plgrid/libs/mkl/2017.0.0 loaded.
  70.  
  71. The following have been reloaded with a version change:
  72.   1) plgrid/libs/mkl/11.3.1 => plgrid/libs/mkl/2017.0.0
  73.  
  74.  plgrid/libs/mkl/11.3.1 unloaded.
  75.  plgrid/libs/mkl/2017.0.0 loaded.
  76.  
  77. The following have been reloaded with a version change:
  78.   1) plgrid/libs/mkl/11.3.1 => plgrid/libs/mkl/2017.0.0
  79.  
  80.  tools/gcc/6.2.0 loaded.
  81.  tools/gcc/6.2.0 loaded.
  82. PROGRAM_ARGS:  --mkl 0 --dummy 0 --sync 0 --cpu 1 --artificial_slowdown 0 --queue_size 1 --my_sim_master_queue 1 --train_log_path /net/archive/groups/plggluna/adam/experiments/breakout_1517329773.87/storage//atari_trainlog/ --predict_batch_size 16 --dummy_predictor 0 --do_train 1 --simulator_procs 10 --env Breakout-v0 --nr_towers 1 --nr_predict_towers 3 --steps_per_epoch 1000 --fc_neurons 256 --batch_size 128 --learning_rate 0.00015 --port 61216 --tf_port 30065 --optimizer adam --use_sync_opt 1 --num_grad 60 --early_stopping None --ps 1 --fc_init uniform --conv_init normal --replace_with_conv True --fc_splits 1 --debug_charts False --epsilon 1e-08 --beta1 0.9 --beta2 0.999 --save_every 0 --models_dir /net/archive/groups/plggluna/adam/experiments/breakout_1517329773.87/models/ --experiment_dir /net/archive/groups/plggluna/adam/experiments/breakout_1517329773.87 --adam_debug False --eval_node False --record_node False --schedule_hyper False
  83. OFFLINE: False
  84. PROGRAM_ARGS:  --mkl 0 --dummy 0 --sync 0 --cpu 1 --artificial_slowdown 0 --queue_size 1 --my_sim_master_queue 1 --train_log_path /net/archive/groups/plggluna/adam/experiments/breakout_1517329773.87/storage//atari_trainlog/ --predict_batch_size 16 --dummy_predictor 0 --do_train 1 --simulator_procs 10 --env Breakout-v0 --nr_towers 1 --nr_predict_towers 3 --steps_per_epoch 1000 --fc_neurons 256 --batch_size 128 --learning_rate 0.00015 --port 61216 --tf_port 30065 --optimizer adam --use_sync_opt 1 --num_grad 60 --early_stopping None --ps 1 --fc_init uniform --conv_init normal --replace_with_conv True --fc_splits 1 --debug_charts False --epsilon 1e-08 --beta1 0.9 --beta2 0.999 --save_every 0 --models_dir /net/archive/groups/plggluna/adam/experiments/breakout_1517329773.87/models/ --experiment_dir /net/archive/groups/plggluna/adam/experiments/breakout_1517329773.87 --adam_debug False --eval_node False --record_node False --schedule_hyper False
  85. OFFLINE: False
  86. PROGRAM_ARGS:  --mkl 0 --dummy 0 --sync 0 --cpu 1 --artificial_slowdown 0 --queue_size 1 --my_sim_master_queue 1 --train_log_path /net/archive/groups/plggluna/adam/experiments/breakout_1517329773.87/storage//atari_trainlog/ --predict_batch_size 16 --dummy_predictor 0 --do_train 1 --simulator_procs 10 --env Breakout-v0 --nr_towers 1 --nr_predict_towers 3 --steps_per_epoch 1000 --fc_neurons 256 --batch_size 128 --learning_rate 0.00015 --port 61216 --tf_port 30065 --optimizer adam --use_sync_opt 1 --num_grad 60 --early_stopping None --ps 1 --fc_init uniform --conv_init normal --replace_with_conv True --fc_splits 1 --debug_charts False --epsilon 1e-08 --beta1 0.9 --beta2 0.999 --save_every 0 --models_dir /net/archive/groups/plggluna/adam/experiments/breakout_1517329773.87/models/ --experiment_dir /net/archive/groups/plggluna/adam/experiments/breakout_1517329773.87 --adam_debug False --eval_node False --record_node False --schedule_hyper False
  87. OFFLINE: False
  88. 2018-01-30 17:29:48.435736: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.1 instructions, but these are available on your machine and could speed up CPU computations.
  89. 2018-01-30 17:29:48.435772: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.2 instructions, but these are available on your machine and could speed up CPU computations.
  90. 2018-01-30 17:29:48.435780: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX instructions, but these are available on your machine and could speed up CPU computations.
  91. 2018-01-30 17:29:48.435788: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX2 instructions, but these are available on your machine and could speed up CPU computations.
  92. 2018-01-30 17:29:48.435795: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use FMA instructions, but these are available on your machine and could speed up CPU computations.
  93. 2018-01-30 17:29:48.445459: I tensorflow/core/distributed_runtime/rpc/grpc_channel.cc:215] Initialize GrpcChannelCache for job ps -> {0 -> localhost:30065}
  94. 2018-01-30 17:29:48.445499: I tensorflow/core/distributed_runtime/rpc/grpc_channel.cc:215] Initialize GrpcChannelCache for job worker -> {0 -> p0112:30066, 1 -> p0115:30066, 2 -> p0574:30066, 3 -> p0576:30066}
  95. 2018-01-30 17:29:48.446966: I tensorflow/core/distributed_runtime/rpc/grpc_server_lib.cc:316] Started server with target: grpc://localhost:30065
  96. {'ps': ['p0111:30065'], 'worker': ['p0112:30066', 'p0115:30066', 'p0574:30066', 'p0576:30066']}
  97. [ps:0] Starting the TF server
  98. [0130 17:29:48 @train.py:84] [ps:0] joining the server.
  99. 2018-01-30 17:29:48.492019: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.1 instructions, but these are available on your machine and could speed up CPU computations.
  100. 2018-01-30 17:29:48.492076: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.2 instructions, but these are available on your machine and could speed up CPU computations.
  101. 2018-01-30 17:29:48.492084: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX instructions, but these are available on your machine and could speed up CPU computations.
  102. 2018-01-30 17:29:48.492091: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX2 instructions, but these are available on your machine and could speed up CPU computations.
  103. 2018-01-30 17:29:48.492097: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use FMA instructions, but these are available on your machine and could speed up CPU computations.
  104. 2018-01-30 17:29:48.502455: I tensorflow/core/distributed_runtime/rpc/grpc_channel.cc:215] Initialize GrpcChannelCache for job ps -> {0 -> p0111:30065}
  105. 2018-01-30 17:29:48.502505: I tensorflow/core/distributed_runtime/rpc/grpc_channel.cc:215] Initialize GrpcChannelCache for job worker -> {0 -> localhost:30066, 1 -> p0115:30066, 2 -> p0574:30066, 3 -> p0576:30066}
  106. 2018-01-30 17:29:48.504127: I tensorflow/core/distributed_runtime/rpc/grpc_server_lib.cc:316] Started server with target: grpc://localhost:30066
  107. [2018-01-30 17:29:48,505] Making new env: Breakout-v0
  108. {'ps': ['p0111:30065'], 'worker': ['p0112:30066', 'p0115:30066', 'p0574:30066', 'p0576:30066']}
  109. [worker:0] Starting the TF server
  110. args.mkl ==  0
  111. using tensorflow convolution
  112. [2018-01-30 17:29:48,609] Making new env: Breakout-v0
  113. [2018-01-30 17:29:48,614] Making new env: Breakout-v0
  114. [2018-01-30 17:29:48,620] Making new env: Breakout-v0
  115. [2018-01-30 17:29:48,625] Making new env: Breakout-v0
  116. [2018-01-30 17:29:48,631] Making new env: Breakout-v0
  117. [2018-01-30 17:29:48,637] Making new env: Breakout-v0
  118. [2018-01-30 17:29:48,643] Making new env: Breakout-v0
  119. [2018-01-30 17:29:48,649] Making new env: Breakout-v0
  120. [2018-01-30 17:29:48,655] Making new env: Breakout-v0
  121. [2018-01-30 17:29:48,661] Making new env: Breakout-v0
  122. None <type 'NoneType'>
  123.  
  124.  
  125.  worker host: grpc://localhost:30066
  126.  
  127.  
  128. [0130 17:29:48 @train.py:711] [BA3C] Train on gpu 0 and infer on gpu 0,0,0
  129. [0130 17:29:48 @train.py:717] using async version
  130. DUMMY PREDICTOR 0
  131. MultiGPUTrainer __init__ dummy = 0
  132. [0130 17:29:48 @multigpu.py:57] Training a model of 1 tower
  133. [0130 17:29:48 @multigpu.py:67] Building graph for training tower 0..., /cpu:0
  134. 12
  135. [0130 17:29:48 @_common.py:61] conv0 input: [None, 84, 84, 16]
  136. Tensor("tower0/conv0/output:0", shape=(?, 80, 80, 32), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  137. [0130 17:29:48 @_common.py:70] conv0 output: [None, 80, 80, 32]
  138. [0130 17:29:48 @_common.py:61] pool0 input: [None, 80, 80, 32]
  139. Tensor("tower0/pool0/MaxPool:0", shape=(?, 40, 40, 32), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  140. [0130 17:29:48 @_common.py:70] pool0 output: [None, 40, 40, 32]
  141. [0130 17:29:48 @_common.py:61] conv1 input: [None, 40, 40, 32]
  142. Tensor("tower0/conv1/output:0", shape=(?, 36, 36, 32), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  143. [0130 17:29:48 @_common.py:70] conv1 output: [None, 36, 36, 32]
  144. [0130 17:29:48 @_common.py:61] pool1 input: [None, 36, 36, 32]
  145. Tensor("tower0/pool1/MaxPool:0", shape=(?, 18, 18, 32), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  146. [0130 17:29:48 @_common.py:70] pool1 output: [None, 18, 18, 32]
  147. [0130 17:29:48 @_common.py:61] conv2 input: [None, 18, 18, 32]
  148. Tensor("tower0/conv2/output:0", shape=(?, 14, 14, 64), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  149. [0130 17:29:48 @_common.py:70] conv2 output: [None, 14, 14, 64]
  150. [0130 17:29:48 @_common.py:61] pool2 input: [None, 14, 14, 64]
  151. Tensor("tower0/pool2/MaxPool:0", shape=(?, 7, 7, 64), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  152. [0130 17:29:48 @_common.py:70] pool2 output: [None, 7, 7, 64]
  153. [0130 17:29:48 @_common.py:61] conv3 input: [None, 7, 7, 64]
  154. Tensor("tower0/conv3/output:0", shape=(?, 5, 5, 64), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  155. [0130 17:29:48 @_common.py:70] conv3 output: [None, 5, 5, 64]
  156. [0130 17:29:48 @_common.py:61] fc1_0 input: [None, 5, 5, 64]
  157. Tensor("tower0/fc1_0/output:0", shape=(?, 1, 1, 256), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  158. [0130 17:29:48 @_common.py:70] fc1_0 output: [None, 1, 1, 256]
  159. [0130 17:29:48 @_common.py:61] fc-pi input: [None, 256]
  160. Tensor("tower0/fc-pi/output:0", shape=(?, 6), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  161. [0130 17:29:48 @_common.py:70] fc-pi output: [None, 6]
  162. [0130 17:29:48 @_common.py:61] fc-v input: [None, 256]
  163. PROGRAM_ARGS:  --mkl 0 --dummy 0 --sync 0 --cpu 1 --artificial_slowdown 0 --queue_size 1 --my_sim_master_queue 1 --train_log_path /net/archive/groups/plggluna/adam/experiments/breakout_1517329773.87/storage//atari_trainlog/ --predict_batch_size 16 --dummy_predictor 0 --do_train 1 --simulator_procs 10 --env Breakout-v0 --nr_towers 1 --nr_predict_towers 3 --steps_per_epoch 1000 --fc_neurons 256 --batch_size 128 --learning_rate 0.00015 --port 61216 --tf_port 30065 --optimizer adam --use_sync_opt 1 --num_grad 60 --early_stopping None --ps 1 --fc_init uniform --conv_init normal --replace_with_conv True --fc_splits 1 --debug_charts False --epsilon 1e-08 --beta1 0.9 --beta2 0.999 --save_every 0 --models_dir /net/archive/groups/plggluna/adam/experiments/breakout_1517329773.87/models/ --experiment_dir /net/archive/groups/plggluna/adam/experiments/breakout_1517329773.87 --adam_debug False --eval_node False --record_node False --schedule_hyper False
  164. OFFLINE: False
  165. PROGRAM_ARGS:  --mkl 0 --dummy 0 --sync 0 --cpu 1 --artificial_slowdown 0 --queue_size 1 --my_sim_master_queue 1 --train_log_path /net/archive/groups/plggluna/adam/experiments/breakout_1517329773.87/storage//atari_trainlog/ --predict_batch_size 16 --dummy_predictor 0 --do_train 1 --simulator_procs 10 --env Breakout-v0 --nr_towers 1 --nr_predict_towers 3 --steps_per_epoch 1000 --fc_neurons 256 --batch_size 128 --learning_rate 0.00015 --port 61216 --tf_port 30065 --optimizer adam --use_sync_opt 1 --num_grad 60 --early_stopping None --ps 1 --fc_init uniform --conv_init normal --replace_with_conv True --fc_splits 1 --debug_charts False --epsilon 1e-08 --beta1 0.9 --beta2 0.999 --save_every 0 --models_dir /net/archive/groups/plggluna/adam/experiments/breakout_1517329773.87/models/ --experiment_dir /net/archive/groups/plggluna/adam/experiments/breakout_1517329773.87 --adam_debug False --eval_node False --record_node False --schedule_hyper False
  166. OFFLINE: False
  167. 2018-01-30 17:29:48.970453: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.1 instructions, but these are available on your machine and could speed up CPU computations.
  168. 2018-01-30 17:29:48.970511: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.2 instructions, but these are available on your machine and could speed up CPU computations.
  169. 2018-01-30 17:29:48.970531: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX instructions, but these are available on your machine and could speed up CPU computations.
  170. 2018-01-30 17:29:48.970539: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX2 instructions, but these are available on your machine and could speed up CPU computations.
  171. 2018-01-30 17:29:48.970546: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use FMA instructions, but these are available on your machine and could speed up CPU computations.
  172. 2018-01-30 17:29:48.981347: I tensorflow/core/distributed_runtime/rpc/grpc_channel.cc:215] Initialize GrpcChannelCache for job ps -> {0 -> p0111:30065}
  173. 2018-01-30 17:29:48.981385: I tensorflow/core/distributed_runtime/rpc/grpc_channel.cc:215] Initialize GrpcChannelCache for job worker -> {0 -> p0112:30066, 1 -> localhost:30066, 2 -> p0574:30066, 3 -> p0576:30066}
  174. 2018-01-30 17:29:48.983013: I tensorflow/core/distributed_runtime/rpc/grpc_server_lib.cc:316] Started server with target: grpc://localhost:30066
  175. [2018-01-30 17:29:48,984] Making new env: Breakout-v0
  176. {'ps': ['p0111:30065'], 'worker': ['p0112:30066', 'p0115:30066', 'p0574:30066', 'p0576:30066']}
  177. [worker:1] Starting the TF server
  178. args.mkl ==  0
  179. Tensor("tower0/fc-v/output:0", shape=(?, 1), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  180. [0130 17:29:49 @_common.py:70] fc-v output: [None, 1]
  181. using tensorflow convolution
  182. [2018-01-30 17:29:49,097] Making new env: Breakout-v0
  183. [2018-01-30 17:29:49,103] Making new env: Breakout-v0
  184. [2018-01-30 17:29:49,109] Making new env: Breakout-v0
  185. [2018-01-30 17:29:49,114] Making new env: Breakout-v0
  186. [2018-01-30 17:29:49,120] Making new env: Breakout-v0
  187. [2018-01-30 17:29:49,126] Making new env: Breakout-v0
  188. [2018-01-30 17:29:49,132] Making new env: Breakout-v0
  189. [2018-01-30 17:29:49,138] Making new env: Breakout-v0
  190. [2018-01-30 17:29:49,145] Making new env: Breakout-v0
  191. [2018-01-30 17:29:49,151] Making new env: Breakout-v0
  192. None <type 'NoneType'>
  193.  
  194.  
  195.  worker host: grpc://localhost:30066
  196.  
  197.  
  198. [0130 17:29:49 @train.py:711] [BA3C] Train on gpu 0 and infer on gpu 0,0,0
  199. [0130 17:29:49 @train.py:717] using async version
  200. DUMMY PREDICTOR 0
  201. MultiGPUTrainer __init__ dummy = 0
  202. [0130 17:29:49 @multigpu.py:57] Training a model of 1 tower
  203. [0130 17:29:49 @multigpu.py:67] Building graph for training tower 0..., /cpu:0
  204. 12
  205. [0130 17:29:49 @_common.py:61] conv0 input: [None, 84, 84, 16]
  206. Tensor("tower0/conv0/output:0", shape=(?, 80, 80, 32), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  207. [0130 17:29:49 @_common.py:70] conv0 output: [None, 80, 80, 32]
  208. [0130 17:29:49 @_common.py:61] pool0 input: [None, 80, 80, 32]
  209. Tensor("tower0/pool0/MaxPool:0", shape=(?, 40, 40, 32), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  210. [0130 17:29:49 @_common.py:70] pool0 output: [None, 40, 40, 32]
  211. [0130 17:29:49 @_common.py:61] conv1 input: [None, 40, 40, 32]
  212. Tensor("tower0/conv1/output:0", shape=(?, 36, 36, 32), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  213. [0130 17:29:49 @_common.py:70] conv1 output: [None, 36, 36, 32]
  214. [0130 17:29:49 @_common.py:61] pool1 input: [None, 36, 36, 32]
  215. Tensor("tower0/pool1/MaxPool:0", shape=(?, 18, 18, 32), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  216. [0130 17:29:49 @_common.py:70] pool1 output: [None, 18, 18, 32]
  217. [0130 17:29:49 @_common.py:61] conv2 input: [None, 18, 18, 32]
  218. Tensor("tower0/conv2/output:0", shape=(?, 14, 14, 64), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  219. [0130 17:29:49 @_common.py:70] conv2 output: [None, 14, 14, 64]
  220. [0130 17:29:49 @_common.py:61] pool2 input: [None, 14, 14, 64]
  221. Tensor("tower0/pool2/MaxPool:0", shape=(?, 7, 7, 64), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  222. [0130 17:29:49 @_common.py:70] pool2 output: [None, 7, 7, 64]
  223. [0130 17:29:49 @_common.py:61] conv3 input: [None, 7, 7, 64]
  224. Tensor("tower0/conv3/output:0", shape=(?, 5, 5, 64), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  225. [0130 17:29:49 @_common.py:70] conv3 output: [None, 5, 5, 64]
  226. [0130 17:29:49 @_common.py:61] fc1_0 input: [None, 5, 5, 64]
  227. Tensor("tower0/fc1_0/output:0", shape=(?, 1, 1, 256), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  228. [0130 17:29:49 @_common.py:70] fc1_0 output: [None, 1, 1, 256]
  229. [0130 17:29:49 @_common.py:61] fc-pi input: [None, 256]
  230. Tensor("tower0/fc-pi/output:0", shape=(?, 6), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  231. [0130 17:29:49 @_common.py:70] fc-pi output: [None, 6]
  232. [0130 17:29:49 @_common.py:61] fc-v input: [None, 256]
  233. Tensor("tower0/fc-v/output:0", shape=(?, 1), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  234. [0130 17:29:49 @_common.py:70] fc-v output: [None, 1]
  235. MOVING_SUMMARY_VARIABLES
  236. []
  237. [0130 17:29:49 @modelutils.py:22] Model Parameters:
  238. conv0/W:0: shape=[5, 5, 16, 32], dim=12800
  239. conv1/W:0: shape=[5, 5, 32, 32], dim=25600
  240. conv2/W:0: shape=[5, 5, 32, 64], dim=51200
  241. conv3/W:0: shape=[3, 3, 64, 64], dim=36864
  242. fc1_0/W:0: shape=[5, 5, 64, 256], dim=409600
  243. fc-pi/W:0: shape=[256, 6], dim=1536
  244. fc-pi/b:0: shape=[6], dim=6
  245. fc-v/W:0: shape=[256, 1], dim=256
  246. fc-v/b:0: shape=[1], dim=1
  247. Total param=537863 (2.051785 MB assuming all float32)
  248. MOVING_SUMMARY_VARIABLES
  249. []
  250. [0130 17:29:50 @modelutils.py:22] Model Parameters:
  251. conv0/W:0: shape=[5, 5, 16, 32], dim=12800
  252. conv1/W:0: shape=[5, 5, 32, 32], dim=25600
  253. conv2/W:0: shape=[5, 5, 32, 64], dim=51200
  254. conv3/W:0: shape=[3, 3, 64, 64], dim=36864
  255. fc1_0/W:0: shape=[5, 5, 64, 256], dim=409600
  256. fc-pi/W:0: shape=[256, 6], dim=1536
  257. fc-pi/b:0: shape=[6], dim=6
  258. fc-v/W:0: shape=[256, 1], dim=256
  259. fc-v/b:0: shape=[1], dim=1
  260. Total param=537863 (2.051785 MB assuming all float32)
  261. [0130 17:29:50 @multigpu.py:228] Setup callbacks ...
  262. Creating Predictorfactor 0
  263. [0130 17:29:50 @base.py:132] Building graph for predictor tower 0...
  264. 12
  265. Tensor("towerp0/conv0/output:0", shape=(?, 80, 80, 32), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  266. Tensor("towerp0/pool0/MaxPool:0", shape=(?, 40, 40, 32), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  267. Tensor("towerp0/conv1/output:0", shape=(?, 36, 36, 32), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  268. Tensor("towerp0/pool1/MaxPool:0", shape=(?, 18, 18, 32), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  269. Tensor("towerp0/conv2/output:0", shape=(?, 14, 14, 64), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  270. Tensor("towerp0/pool2/MaxPool:0", shape=(?, 7, 7, 64), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  271. Tensor("towerp0/conv3/output:0", shape=(?, 5, 5, 64), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  272. Tensor("towerp0/fc1_0/output:0", shape=(?, 1, 1, 256), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  273. Tensor("towerp0/fc-pi/output:0", shape=(?, 6), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  274. Tensor("towerp0/fc-v/output:0", shape=(?, 1), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  275. [0130 17:29:51 @base.py:132] Building graph for predictor tower 0...
  276. 12
  277. Tensor("towerp0_1/conv0/output:0", shape=(?, 80, 80, 32), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  278. Tensor("towerp0_1/pool0/MaxPool:0", shape=(?, 40, 40, 32), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  279. Tensor("towerp0_1/conv1/output:0", shape=(?, 36, 36, 32), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  280. Tensor("towerp0_1/pool1/MaxPool:0", shape=(?, 18, 18, 32), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  281. Tensor("towerp0_1/conv2/output:0", shape=(?, 14, 14, 64), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  282. Tensor("towerp0_1/pool2/MaxPool:0", shape=(?, 7, 7, 64), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  283. Tensor("towerp0_1/conv3/output:0", shape=(?, 5, 5, 64), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  284. Tensor("towerp0_1/fc1_0/output:0", shape=(?, 1, 1, 256), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  285. Tensor("towerp0_1/fc-pi/output:0", shape=(?, 6), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  286. Tensor("towerp0_1/fc-v/output:0", shape=(?, 1), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  287. [0130 17:29:51 @base.py:132] Building graph for predictor tower 0...
  288. 12
  289. Tensor("towerp0_2/conv0/output:0", shape=(?, 80, 80, 32), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  290. Tensor("towerp0_2/pool0/MaxPool:0", shape=(?, 40, 40, 32), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  291. Tensor("towerp0_2/conv1/output:0", shape=(?, 36, 36, 32), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  292. Tensor("towerp0_2/pool1/MaxPool:0", shape=(?, 18, 18, 32), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  293. Tensor("towerp0_2/conv2/output:0", shape=(?, 14, 14, 64), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  294. Tensor("towerp0_2/pool2/MaxPool:0", shape=(?, 7, 7, 64), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  295. Tensor("towerp0_2/conv3/output:0", shape=(?, 5, 5, 64), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  296. Tensor("towerp0_2/fc1_0/output:0", shape=(?, 1, 1, 256), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  297. Tensor("towerp0_2/fc-pi/output:0", shape=(?, 6), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  298. Tensor("towerp0_2/fc-v/output:0", shape=(?, 1), dtype=float32, device=/job:worker/task:0/device:CPU:0)
  299. [0130 17:29:51 @base.py:177] ===============================================================
  300. [0130 17:29:51 @base.py:179] CHIEF!
  301. [0130 17:29:51 @base.py:180] [p0112] Creating the session
  302. [0130 17:29:51 @base.py:181] ===============================================================
  303. [0130 17:29:51 @multigpu.py:228] Setup callbacks ...
  304. Creating Predictorfactor 0
  305. [0130 17:29:51 @base.py:132] Building graph for predictor tower 0...
  306. 12
  307. Tensor("towerp0/conv0/output:0", shape=(?, 80, 80, 32), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  308. Tensor("towerp0/pool0/MaxPool:0", shape=(?, 40, 40, 32), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  309. Tensor("towerp0/conv1/output:0", shape=(?, 36, 36, 32), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  310. Tensor("towerp0/pool1/MaxPool:0", shape=(?, 18, 18, 32), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  311. Tensor("towerp0/conv2/output:0", shape=(?, 14, 14, 64), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  312. Tensor("towerp0/pool2/MaxPool:0", shape=(?, 7, 7, 64), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  313. Tensor("towerp0/conv3/output:0", shape=(?, 5, 5, 64), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  314. Tensor("towerp0/fc1_0/output:0", shape=(?, 1, 1, 256), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  315. Tensor("towerp0/fc-pi/output:0", shape=(?, 6), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  316. Tensor("towerp0/fc-v/output:0", shape=(?, 1), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  317. [0130 17:29:51 @base.py:132] Building graph for predictor tower 0...
  318. 12
  319. Tensor("towerp0_1/conv0/output:0", shape=(?, 80, 80, 32), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  320. Tensor("towerp0_1/pool0/MaxPool:0", shape=(?, 40, 40, 32), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  321. Tensor("towerp0_1/conv1/output:0", shape=(?, 36, 36, 32), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  322. Tensor("towerp0_1/pool1/MaxPool:0", shape=(?, 18, 18, 32), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  323. Tensor("towerp0_1/conv2/output:0", shape=(?, 14, 14, 64), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  324. Tensor("towerp0_1/pool2/MaxPool:0", shape=(?, 7, 7, 64), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  325. Tensor("towerp0_1/conv3/output:0", shape=(?, 5, 5, 64), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  326. Tensor("towerp0_1/fc1_0/output:0", shape=(?, 1, 1, 256), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  327. Tensor("towerp0_1/fc-pi/output:0", shape=(?, 6), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  328. Tensor("towerp0_1/fc-v/output:0", shape=(?, 1), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  329. [0130 17:29:51 @base.py:132] Building graph for predictor tower 0...
  330. 12
  331. Tensor("towerp0_2/conv0/output:0", shape=(?, 80, 80, 32), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  332. Tensor("towerp0_2/pool0/MaxPool:0", shape=(?, 40, 40, 32), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  333. Tensor("towerp0_2/conv1/output:0", shape=(?, 36, 36, 32), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  334. Tensor("towerp0_2/pool1/MaxPool:0", shape=(?, 18, 18, 32), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  335. Tensor("towerp0_2/conv2/output:0", shape=(?, 14, 14, 64), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  336. Tensor("towerp0_2/pool2/MaxPool:0", shape=(?, 7, 7, 64), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  337. Tensor("towerp0_2/conv3/output:0", shape=(?, 5, 5, 64), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  338. Tensor("towerp0_2/fc1_0/output:0", shape=(?, 1, 1, 256), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  339. Tensor("towerp0_2/fc-pi/output:0", shape=(?, 6), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  340. Tensor("towerp0_2/fc-v/output:0", shape=(?, 1), dtype=float32, device=/job:worker/task:1/device:CPU:0)
  341. [0130 17:29:51 @base.py:177] ===============================================================
  342. [0130 17:29:51 @base.py:180] [p0115] Creating the session
  343. [0130 17:29:51 @base.py:181] ===============================================================
  344. 2018-01-30 17:29:57.665912: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.1 instructions, but these are available on your machine and could speed up CPU computations.
  345. 2018-01-30 17:29:57.665949: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.2 instructions, but these are available on your machine and could speed up CPU computations.
  346. 2018-01-30 17:29:57.665958: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX instructions, but these are available on your machine and could speed up CPU computations.
  347. 2018-01-30 17:29:57.665965: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX2 instructions, but these are available on your machine and could speed up CPU computations.
  348. 2018-01-30 17:29:57.665973: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use FMA instructions, but these are available on your machine and could speed up CPU computations.
  349. 2018-01-30 17:29:57.670158: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.1 instructions, but these are available on your machine and could speed up CPU computations.
  350. 2018-01-30 17:29:57.670198: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.2 instructions, but these are available on your machine and could speed up CPU computations.
  351. 2018-01-30 17:29:57.670207: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX instructions, but these are available on your machine and could speed up CPU computations.
  352. 2018-01-30 17:29:57.670215: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX2 instructions, but these are available on your machine and could speed up CPU computations.
  353. 2018-01-30 17:29:57.670222: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use FMA instructions, but these are available on your machine and could speed up CPU computations.
  354. 2018-01-30 17:29:57.676919: I tensorflow/core/distributed_runtime/rpc/grpc_channel.cc:215] Initialize GrpcChannelCache for job ps -> {0 -> p0111:30065}
  355. 2018-01-30 17:29:57.676959: I tensorflow/core/distributed_runtime/rpc/grpc_channel.cc:215] Initialize GrpcChannelCache for job worker -> {0 -> p0112:30066, 1 -> p0115:30066, 2 -> p0574:30066, 3 -> localhost:30066}
  356. 2018-01-30 17:29:57.678659: I tensorflow/core/distributed_runtime/rpc/grpc_server_lib.cc:316] Started server with target: grpc://localhost:30066
  357. 2018-01-30 17:29:57.680658: I tensorflow/core/distributed_runtime/rpc/grpc_channel.cc:215] Initialize GrpcChannelCache for job ps -> {0 -> p0111:30065}
  358. [2018-01-30 17:29:57,679] Making new env: Breakout-v0
  359. 2018-01-30 17:29:57.680702: I tensorflow/core/distributed_runtime/rpc/grpc_channel.cc:215] Initialize GrpcChannelCache for job worker -> {0 -> p0112:30066, 1 -> p0115:30066, 2 -> localhost:30066, 3 -> p0576:30066}
  360. 2018-01-30 17:29:57.682393: I tensorflow/core/distributed_runtime/rpc/grpc_server_lib.cc:316] Started server with target: grpc://localhost:30066
  361. [2018-01-30 17:29:57,683] Making new env: Breakout-v0
  362. {'ps': ['p0111:30065'], 'worker': ['p0112:30066', 'p0115:30066', 'p0574:30066', 'p0576:30066']}
  363. [worker:3] Starting the TF server
  364. args.mkl ==  0
  365. {'ps': ['p0111:30065'], 'worker': ['p0112:30066', 'p0115:30066', 'p0574:30066', 'p0576:30066']}
  366. [worker:2] Starting the TF server
  367. args.mkl ==  0
  368. using tensorflow convolution
  369. using tensorflow convolution
  370. [2018-01-30 17:29:57,918] Making new env: Breakout-v0
  371. [2018-01-30 17:29:57,917] Making new env: Breakout-v0
  372. [2018-01-30 17:29:57,922] Making new env: Breakout-v0
  373. [2018-01-30 17:29:57,923] Making new env: Breakout-v0
  374. [2018-01-30 17:29:57,928] Making new env: Breakout-v0
  375. [2018-01-30 17:29:57,929] Making new env: Breakout-v0
  376. [2018-01-30 17:29:57,934] Making new env: Breakout-v0
  377. [2018-01-30 17:29:57,935] Making new env: Breakout-v0
  378. [2018-01-30 17:29:57,939] Making new env: Breakout-v0
  379. [2018-01-30 17:29:57,940] Making new env: Breakout-v0
  380. [2018-01-30 17:29:57,952] Making new env: Breakout-v0
  381. [2018-01-30 17:29:57,956] Making new env: Breakout-v0
  382. [2018-01-30 17:29:57,957] Making new env: Breakout-v0
  383. [2018-01-30 17:29:57,959] Making new env: Breakout-v0
  384. [2018-01-30 17:29:57,961] Making new env: Breakout-v0
  385. [2018-01-30 17:29:57,964] Making new env: Breakout-v0
  386. [2018-01-30 17:29:57,965] Making new env: Breakout-v0
  387. [2018-01-30 17:29:57,971] Making new env: Breakout-v0
  388. [2018-01-30 17:29:57,978] Making new env: Breakout-v0
  389. [2018-01-30 17:29:57,978] Making new env: Breakout-v0
  390. None <type 'NoneType'>
  391.  
  392.  
  393.  worker host: grpc://localhost:30066
  394.  
  395.  
  396. [0130 17:29:57 @train.py:711] [BA3C] Train on gpu 0 and infer on gpu 0,0,0
  397. [0130 17:29:57 @train.py:717] using async version
  398. DUMMY PREDICTOR 0
  399. MultiGPUTrainer __init__ dummy = 0
  400. [0130 17:29:57 @multigpu.py:57] Training a model of 1 tower
  401. [0130 17:29:57 @multigpu.py:67] Building graph for training tower 0..., /cpu:0
  402. None <type 'NoneType'>
  403.  
  404.  
  405.  worker host: grpc://localhost:30066
  406.  
  407.  
  408. [0130 17:29:58 @train.py:711] [BA3C] Train on gpu 0 and infer on gpu 0,0,0
  409. [0130 17:29:58 @train.py:717] using async version
  410. 12
  411. DUMMY PREDICTOR 0
  412. MultiGPUTrainer __init__ dummy = 0
  413. [0130 17:29:58 @multigpu.py:57] Training a model of 1 tower
  414. [0130 17:29:58 @multigpu.py:67] Building graph for training tower 0..., /cpu:0
  415. [0130 17:29:58 @_common.py:61] conv0 input: [None, 84, 84, 16]
  416. 12
  417. Tensor("tower0/conv0/output:0", shape=(?, 80, 80, 32), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  418. [0130 17:29:58 @_common.py:70] conv0 output: [None, 80, 80, 32]
  419. [0130 17:29:58 @_common.py:61] pool0 input: [None, 80, 80, 32]
  420. Tensor("tower0/pool0/MaxPool:0", shape=(?, 40, 40, 32), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  421. [0130 17:29:58 @_common.py:70] pool0 output: [None, 40, 40, 32]
  422. [0130 17:29:58 @_common.py:61] conv1 input: [None, 40, 40, 32]
  423. [0130 17:29:58 @_common.py:61] conv0 input: [None, 84, 84, 16]
  424. Tensor("tower0/conv1/output:0", shape=(?, 36, 36, 32), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  425. [0130 17:29:58 @_common.py:70] conv1 output: [None, 36, 36, 32]
  426. [0130 17:29:58 @_common.py:61] pool1 input: [None, 36, 36, 32]
  427. Tensor("tower0/pool1/MaxPool:0", shape=(?, 18, 18, 32), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  428. [0130 17:29:58 @_common.py:70] pool1 output: [None, 18, 18, 32]
  429. [0130 17:29:58 @_common.py:61] conv2 input: [None, 18, 18, 32]
  430. Tensor("tower0/conv0/output:0", shape=(?, 80, 80, 32), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  431. [0130 17:29:58 @_common.py:70] conv0 output: [None, 80, 80, 32]
  432. [0130 17:29:58 @_common.py:61] pool0 input: [None, 80, 80, 32]
  433. Tensor("tower0/conv2/output:0", shape=(?, 14, 14, 64), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  434. [0130 17:29:58 @_common.py:70] conv2 output: [None, 14, 14, 64]
  435. Tensor("tower0/pool0/MaxPool:0", shape=(?, 40, 40, 32), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  436. [0130 17:29:58 @_common.py:70] pool0 output: [None, 40, 40, 32]
  437. [0130 17:29:58 @_common.py:61] conv1 input: [None, 40, 40, 32]
  438. [0130 17:29:58 @_common.py:61] pool2 input: [None, 14, 14, 64]
  439. Tensor("tower0/pool2/MaxPool:0", shape=(?, 7, 7, 64), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  440. [0130 17:29:58 @_common.py:70] pool2 output: [None, 7, 7, 64]
  441. [0130 17:29:58 @_common.py:61] conv3 input: [None, 7, 7, 64]
  442. Tensor("tower0/conv1/output:0", shape=(?, 36, 36, 32), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  443. [0130 17:29:58 @_common.py:70] conv1 output: [None, 36, 36, 32]
  444. Tensor("tower0/conv3/output:0", shape=(?, 5, 5, 64), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  445. [0130 17:29:58 @_common.py:70] conv3 output: [None, 5, 5, 64]
  446. [0130 17:29:58 @_common.py:61] pool1 input: [None, 36, 36, 32]
  447. Tensor("tower0/pool1/MaxPool:0", shape=(?, 18, 18, 32), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  448. [0130 17:29:58 @_common.py:70] pool1 output: [None, 18, 18, 32]
  449. [0130 17:29:58 @_common.py:61] conv2 input: [None, 18, 18, 32]
  450. [0130 17:29:58 @_common.py:61] fc1_0 input: [None, 5, 5, 64]
  451. Tensor("tower0/conv2/output:0", shape=(?, 14, 14, 64), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  452. [0130 17:29:58 @_common.py:70] conv2 output: [None, 14, 14, 64]
  453. Tensor("tower0/fc1_0/output:0", shape=(?, 1, 1, 256), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  454. [0130 17:29:58 @_common.py:70] fc1_0 output: [None, 1, 1, 256]
  455. [0130 17:29:58 @_common.py:61] pool2 input: [None, 14, 14, 64]
  456. Tensor("tower0/pool2/MaxPool:0", shape=(?, 7, 7, 64), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  457. [0130 17:29:58 @_common.py:70] pool2 output: [None, 7, 7, 64]
  458. [0130 17:29:58 @_common.py:61] conv3 input: [None, 7, 7, 64]
  459. [0130 17:29:58 @_common.py:61] fc-pi input: [None, 256]
  460. Tensor("tower0/conv3/output:0", shape=(?, 5, 5, 64), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  461. [0130 17:29:58 @_common.py:70] conv3 output: [None, 5, 5, 64]
  462. [0130 17:29:58 @_common.py:61] fc1_0 input: [None, 5, 5, 64]
  463. Tensor("tower0/fc1_0/output:0", shape=(?, 1, 1, 256), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  464. [0130 17:29:58 @_common.py:70] fc1_0 output: [None, 1, 1, 256]
  465. Tensor("tower0/fc-pi/output:0", shape=(?, 6), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  466. [0130 17:29:58 @_common.py:70] fc-pi output: [None, 6]
  467. [0130 17:29:58 @_common.py:61] fc-v input: [None, 256]
  468. [0130 17:29:58 @_common.py:61] fc-pi input: [None, 256]
  469. Tensor("tower0/fc-pi/output:0", shape=(?, 6), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  470. [0130 17:29:58 @_common.py:70] fc-pi output: [None, 6]
  471. [0130 17:29:58 @_common.py:61] fc-v input: [None, 256]
  472. Tensor("tower0/fc-v/output:0", shape=(?, 1), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  473. [0130 17:29:58 @_common.py:70] fc-v output: [None, 1]
  474. Tensor("tower0/fc-v/output:0", shape=(?, 1), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  475. [0130 17:29:58 @_common.py:70] fc-v output: [None, 1]
  476. MOVING_SUMMARY_VARIABLES
  477. []
  478. [0130 17:29:58 @modelutils.py:22] Model Parameters:
  479. conv0/W:0: shape=[5, 5, 16, 32], dim=12800
  480. conv1/W:0: shape=[5, 5, 32, 32], dim=25600
  481. conv2/W:0: shape=[5, 5, 32, 64], dim=51200
  482. conv3/W:0: shape=[3, 3, 64, 64], dim=36864
  483. fc1_0/W:0: shape=[5, 5, 64, 256], dim=409600
  484. fc-pi/W:0: shape=[256, 6], dim=1536
  485. fc-pi/b:0: shape=[6], dim=6
  486. fc-v/W:0: shape=[256, 1], dim=256
  487. fc-v/b:0: shape=[1], dim=1
  488. Total param=537863 (2.051785 MB assuming all float32)
  489. MOVING_SUMMARY_VARIABLES
  490. []
  491. [0130 17:29:59 @modelutils.py:22] Model Parameters:
  492. conv0/W:0: shape=[5, 5, 16, 32], dim=12800
  493. conv1/W:0: shape=[5, 5, 32, 32], dim=25600
  494. conv2/W:0: shape=[5, 5, 32, 64], dim=51200
  495. conv3/W:0: shape=[3, 3, 64, 64], dim=36864
  496. fc1_0/W:0: shape=[5, 5, 64, 256], dim=409600
  497. fc-pi/W:0: shape=[256, 6], dim=1536
  498. fc-pi/b:0: shape=[6], dim=6
  499. fc-v/W:0: shape=[256, 1], dim=256
  500. fc-v/b:0: shape=[1], dim=1
  501. Total param=537863 (2.051785 MB assuming all float32)
  502. [0130 17:30:00 @multigpu.py:228] Setup callbacks ...
  503. Creating Predictorfactor 0
  504. [0130 17:30:00 @base.py:132] Building graph for predictor tower 0...
  505. 2018-01-30 17:30:00.118372: I tensorflow/core/distributed_runtime/master_session.cc:999] Start master session f133859301d86503 with config:
  506.  
  507. 12
  508. Tensor("towerp0/conv0/output:0", shape=(?, 80, 80, 32), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  509. [0130 17:30:00 @multigpu.py:228] Setup callbacks ...
  510. Tensor("towerp0/pool0/MaxPool:0", shape=(?, 40, 40, 32), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  511. Creating Predictorfactor 0
  512. [0130 17:30:00 @base.py:132] Building graph for predictor tower 0...
  513. Tensor("towerp0/conv1/output:0", shape=(?, 36, 36, 32), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  514. Tensor("towerp0/pool1/MaxPool:0", shape=(?, 18, 18, 32), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  515. Tensor("towerp0/conv2/output:0", shape=(?, 14, 14, 64), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  516. Tensor("towerp0/pool2/MaxPool:0", shape=(?, 7, 7, 64), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  517. 12
  518. Tensor("towerp0/conv3/output:0", shape=(?, 5, 5, 64), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  519. Tensor("towerp0/fc1_0/output:0", shape=(?, 1, 1, 256), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  520. Tensor("towerp0/conv0/output:0", shape=(?, 80, 80, 32), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  521. Tensor("towerp0/fc-pi/output:0", shape=(?, 6), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  522. Tensor("towerp0/pool0/MaxPool:0", shape=(?, 40, 40, 32), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  523. Tensor("towerp0/conv1/output:0", shape=(?, 36, 36, 32), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  524. Tensor("towerp0/pool1/MaxPool:0", shape=(?, 18, 18, 32), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  525. Tensor("towerp0/fc-v/output:0", shape=(?, 1), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  526. Tensor("towerp0/conv2/output:0", shape=(?, 14, 14, 64), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  527. Tensor("towerp0/pool2/MaxPool:0", shape=(?, 7, 7, 64), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  528. Tensor("towerp0/conv3/output:0", shape=(?, 5, 5, 64), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  529. Tensor("towerp0/fc1_0/output:0", shape=(?, 1, 1, 256), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  530. Tensor("towerp0/fc-pi/output:0", shape=(?, 6), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  531. Tensor("towerp0/fc-v/output:0", shape=(?, 1), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  532. [0130 17:30:00 @base.py:132] Building graph for predictor tower 0...
  533. 12
  534. Tensor("towerp0_1/conv0/output:0", shape=(?, 80, 80, 32), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  535. Tensor("towerp0_1/pool0/MaxPool:0", shape=(?, 40, 40, 32), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  536. Tensor("towerp0_1/conv1/output:0", shape=(?, 36, 36, 32), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  537. [0130 17:30:00 @base.py:132] Building graph for predictor tower 0...
  538. Tensor("towerp0_1/pool1/MaxPool:0", shape=(?, 18, 18, 32), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  539. Tensor("towerp0_1/conv2/output:0", shape=(?, 14, 14, 64), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  540. Tensor("towerp0_1/pool2/MaxPool:0", shape=(?, 7, 7, 64), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  541. Tensor("towerp0_1/conv3/output:0", shape=(?, 5, 5, 64), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  542. 12
  543. Tensor("towerp0_1/fc1_0/output:0", shape=(?, 1, 1, 256), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  544. Tensor("towerp0_1/fc-pi/output:0", shape=(?, 6), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  545. Tensor("towerp0_1/conv0/output:0", shape=(?, 80, 80, 32), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  546. Tensor("towerp0_1/pool0/MaxPool:0", shape=(?, 40, 40, 32), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  547. Tensor("towerp0_1/fc-v/output:0", shape=(?, 1), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  548. Tensor("towerp0_1/conv1/output:0", shape=(?, 36, 36, 32), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  549. Tensor("towerp0_1/pool1/MaxPool:0", shape=(?, 18, 18, 32), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  550. Tensor("towerp0_1/conv2/output:0", shape=(?, 14, 14, 64), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  551. Tensor("towerp0_1/pool2/MaxPool:0", shape=(?, 7, 7, 64), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  552. Tensor("towerp0_1/conv3/output:0", shape=(?, 5, 5, 64), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  553. Tensor("towerp0_1/fc1_0/output:0", shape=(?, 1, 1, 256), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  554. [0130 17:30:00 @base.py:132] Building graph for predictor tower 0...
  555. Tensor("towerp0_1/fc-pi/output:0", shape=(?, 6), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  556. Tensor("towerp0_1/fc-v/output:0", shape=(?, 1), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  557. 12
  558. Tensor("towerp0_2/conv0/output:0", shape=(?, 80, 80, 32), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  559. Tensor("towerp0_2/pool0/MaxPool:0", shape=(?, 40, 40, 32), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  560. Tensor("towerp0_2/conv1/output:0", shape=(?, 36, 36, 32), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  561. Tensor("towerp0_2/pool1/MaxPool:0", shape=(?, 18, 18, 32), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  562. Tensor("towerp0_2/conv2/output:0", shape=(?, 14, 14, 64), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  563. Tensor("towerp0_2/pool2/MaxPool:0", shape=(?, 7, 7, 64), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  564. [0130 17:30:00 @base.py:132] Building graph for predictor tower 0...
  565. Tensor("towerp0_2/conv3/output:0", shape=(?, 5, 5, 64), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  566. Tensor("towerp0_2/fc1_0/output:0", shape=(?, 1, 1, 256), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  567. 12
  568. Tensor("towerp0_2/fc-pi/output:0", shape=(?, 6), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  569. Tensor("towerp0_2/fc-v/output:0", shape=(?, 1), dtype=float32, device=/job:worker/task:3/device:CPU:0)
  570. Tensor("towerp0_2/conv0/output:0", shape=(?, 80, 80, 32), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  571. Tensor("towerp0_2/pool0/MaxPool:0", shape=(?, 40, 40, 32), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  572. Tensor("towerp0_2/conv1/output:0", shape=(?, 36, 36, 32), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  573. Tensor("towerp0_2/pool1/MaxPool:0", shape=(?, 18, 18, 32), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  574. Tensor("towerp0_2/conv2/output:0", shape=(?, 14, 14, 64), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  575. Tensor("towerp0_2/pool2/MaxPool:0", shape=(?, 7, 7, 64), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  576. Tensor("towerp0_2/conv3/output:0", shape=(?, 5, 5, 64), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  577. Tensor("towerp0_2/fc1_0/output:0", shape=(?, 1, 1, 256), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  578. [0130 17:30:00 @base.py:177] ===============================================================
  579. [0130 17:30:00 @base.py:180] [p0576] Creating the session
  580. [0130 17:30:00 @base.py:181] ===============================================================
  581. Tensor("towerp0_2/fc-pi/output:0", shape=(?, 6), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  582. Tensor("towerp0_2/fc-v/output:0", shape=(?, 1), dtype=float32, device=/job:worker/task:2/device:CPU:0)
  583. [0130 17:30:00 @base.py:177] ===============================================================
  584. [0130 17:30:00 @base.py:180] [p0574] Creating the session
  585. [0130 17:30:00 @base.py:181] ===============================================================
  586. 2018-01-30 17:30:00.790835: I tensorflow/core/distributed_runtime/master_session.cc:999] Start master session ce42849ba70813b5 with config:
  587.  
  588. [0130 17:30:01 @base.py:189] ===============================================================
  589. [0130 17:30:01 @base.py:190] [p0112] Session created
  590. [0130 17:30:01 @base.py:191] ===============================================================
  591. [0130 17:30:01 @base.py:112] [p0112] Initializing graph variables ...
  592. [0130 17:30:01 @base.py:119] [p0112] Starting concurrency...
  593. [0130 17:30:01 @base.py:198] Starting all threads & procs ...
  594. [0130 17:30:01 @base.py:122] [p0112] Setting default session
  595. [0130 17:30:01 @base.py:125] [p0112] Getting global step
  596. [0130 17:30:01 @base.py:127] [p0112] Start training with global_step=0
  597. server main loop
  598. before socket bind... tcp://*:61216
  599. receiving
  600. 2018-01-30 17:30:01.209105: I tensorflow/core/distributed_runtime/master_session.cc:999] Start master session 2eefc7caccbf1c89 with config:
  601.  
  602. 2018-01-30 17:30:01.276788: I tensorflow/core/distributed_runtime/master_session.cc:999] Start master session e774dcd3d1ac3165 with config:
  603.  
  604. [0130 17:30:01 @base.py:189] ===============================================================
  605. [0130 17:30:01 @base.py:190] [p0576] Session created
  606. [0130 17:30:01 @base.py:191] ===============================================================
  607. [0130 17:30:01 @base.py:112] [p0576] Initializing graph variables ...
  608. [0130 17:30:01 @base.py:119] [p0576] Starting concurrency...
  609. [0130 17:30:01 @base.py:198] Starting all threads & procs ...
  610. [0130 17:30:01 @base.py:122] [p0576] Setting default session
  611. [0130 17:30:01 @base.py:125] [p0576] Getting global step
  612. [0130 17:30:01 @base.py:127] [p0576] Start training with global_step=0
  613. [0130 17:30:01 @base.py:189] ===============================================================
  614. [0130 17:30:01 @base.py:190] [p0574] Session created
  615. [0130 17:30:01 @base.py:191] ===============================================================
  616. [0130 17:30:01 @base.py:112] [p0574] Initializing graph variables ...
  617. [0130 17:30:01 @base.py:119] [p0574] Starting concurrency...
  618. [0130 17:30:01 @base.py:198] Starting all threads & procs ...
  619. [0130 17:30:01 @base.py:122] [p0574] Setting default session
  620. [0130 17:30:01 @base.py:125] [p0574] Getting global step
  621. [0130 17:30:01 @base.py:127] [p0574] Start training with global_step=0
  622. [0130 17:30:03 @multigpu.py:323] [p0576]  step: count(1), step_time 4762.69, mean_step_time 4762.69, it/s 0.21
  623. [0130 17:30:03 @multigpu.py:323] [p0574]  step: count(1), step_time 4870.79, mean_step_time 4870.79, it/s 0.21
  624. [0130 17:30:05 @multigpu.py:323] [p0576]  step: count(2), step_time 1437.01, mean_step_time 3099.85, it/s 0.32
  625. [0130 17:30:05 @multigpu.py:323] [p0574]  step: count(2), step_time 1388.9, mean_step_time 3129.85, it/s 0.32
  626. [0130 17:30:06 @multigpu.py:323] [p0576]  step: count(3), step_time 1392.92, mean_step_time 2530.87, it/s 0.4
  627. [0130 17:30:06 @multigpu.py:323] [p0574]  step: count(3), step_time 1417.79, mean_step_time 2559.16, it/s 0.39
  628. [0130 17:30:07 @multigpu.py:323] [p0576]  step: count(4), step_time 1381.3, mean_step_time 2243.48, it/s 0.45
  629. [0130 17:30:08 @multigpu.py:323] [p0574]  step: count(4), step_time 1380.92, mean_step_time 2264.6, it/s 0.44
  630. [0130 17:30:09 @multigpu.py:323] [p0576]  step: count(5), step_time 1413.79, mean_step_time 2077.54, it/s 0.48
  631. [0130 17:30:09 @multigpu.py:323] [p0574]  step: count(5), step_time 1414.17, mean_step_time 2094.51, it/s 0.48
  632. [0130 17:30:10 @multigpu.py:323] [p0576]  step: count(6), step_time 1387.62, mean_step_time 1962.56, it/s 0.51
  633. [0130 17:30:10 @multigpu.py:323] [p0574]  step: count(6), step_time 1367.55, mean_step_time 1973.35, it/s 0.51
  634. [0130 17:30:12 @multigpu.py:323] [p0576]  step: count(7), step_time 1363.13, mean_step_time 1876.92, it/s 0.53
  635. [0130 17:30:12 @multigpu.py:323] [p0574]  step: count(7), step_time 1377.66, mean_step_time 1888.25, it/s 0.53
  636. [0130 17:30:13 @multigpu.py:323] [p0574]  step: count(8), step_time 1369.67, mean_step_time 1823.43, it/s 0.55
  637. [0130 17:30:13 @multigpu.py:323] [p0576]  step: count(8), step_time 1444.03, mean_step_time 1822.81, it/s 0.55
  638. [0130 17:30:14 @multigpu.py:323] [p0574]  step: count(9), step_time 1376.32, mean_step_time 1773.75, it/s 0.56
  639. [0130 17:30:15 @multigpu.py:323] [p0576]  step: count(9), step_time 1402.34, mean_step_time 1776.09, it/s 0.56
  640. [0130 17:30:16 @multigpu.py:323] [p0574]  step: count(10), step_time 1358.5, mean_step_time 1732.23, it/s 0.58
  641. [0130 17:30:16 @multigpu.py:323] [p0576]  step: count(10), step_time 1372.5, mean_step_time 1735.73, it/s 0.58
  642. [0130 17:30:17 @multigpu.py:323] [p0574]  step: count(11), step_time 1402.19, mean_step_time 1702.22, it/s 0.59
  643. [0130 17:30:17 @multigpu.py:323] [p0576]  step: count(11), step_time 1415.18, mean_step_time 1706.59, it/s 0.59
  644. [0130 17:30:19 @multigpu.py:323] [p0574]  step: count(12), step_time 1360.35, mean_step_time 1673.73, it/s 0.6
  645. [0130 17:30:19 @multigpu.py:323] [p0576]  step: count(12), step_time 1386.91, mean_step_time 1679.95, it/s 0.6
  646. [0130 17:30:20 @multigpu.py:323] [p0574]  step: count(13), step_time 1434.23, mean_step_time 1655.31, it/s 0.6
  647. [0130 17:30:20 @multigpu.py:323] [p0576]  step: count(13), step_time 1391.04, mean_step_time 1657.73, it/s 0.6
  648. [0130 17:30:21 @multigpu.py:323] [p0574]  step: count(14), step_time 1350.67, mean_step_time 1633.55, it/s 0.61
  649. [0130 17:30:22 @multigpu.py:323] [p0576]  step: count(14), step_time 1425.22, mean_step_time 1641.12, it/s 0.61
  650. [0130 17:30:23 @multigpu.py:323] [p0574]  step: count(15), step_time 1427.02, mean_step_time 1619.78, it/s 0.62
  651. [0130 17:30:23 @multigpu.py:323] [p0576]  step: count(15), step_time 1376.45, mean_step_time 1623.48, it/s 0.62
  652. [0130 17:30:24 @multigpu.py:323] [p0574]  step: count(16), step_time 1363.51, mean_step_time 1603.76, it/s 0.62
  653. [0130 17:30:24 @multigpu.py:323] [p0576]  step: count(16), step_time 1441.44, mean_step_time 1612.1, it/s 0.62
  654. [0130 17:30:26 @multigpu.py:323] [p0574]  step: count(17), step_time 1414.29, mean_step_time 1592.62, it/s 0.63
  655. [0130 17:30:26 @multigpu.py:323] [p0576]  step: count(17), step_time 1441.84, mean_step_time 1602.08, it/s 0.62
  656. [0130 17:30:27 @multigpu.py:323] [p0574]  step: count(18), step_time 1425.53, mean_step_time 1583.34, it/s 0.63
  657. [0130 17:30:27 @multigpu.py:323] [p0576]  step: count(18), step_time 1347.63, mean_step_time 1587.95, it/s 0.63
  658. [0130 17:30:28 @multigpu.py:323] [p0574]  step: count(19), step_time 1392.45, mean_step_time 1573.29, it/s 0.64
  659. [0130 17:30:29 @multigpu.py:323] [p0576]  step: count(19), step_time 1404.59, mean_step_time 1578.3, it/s 0.63
  660. [0130 17:30:30 @multigpu.py:323] [p0574]  step: count(20), step_time 1396.35, mean_step_time 1564.44, it/s 0.64
  661. [0130 17:30:30 @multigpu.py:323] [p0576]  step: count(20), step_time 1363.71, mean_step_time 1567.57, it/s 0.64
  662. 2018-01-30 17:30:30.682429: I tensorflow/core/distributed_runtime/master_session.cc:999] Start master session 8454efd2be83293e with config:
  663.  
  664. [0130 17:30:30 @base.py:189] ===============================================================
  665. [0130 17:30:30 @base.py:190] [p0115] Session created
  666. [0130 17:30:30 @base.py:191] ===============================================================
  667. [0130 17:30:30 @base.py:112] [p0115] Initializing graph variables ...
  668. [0130 17:30:30 @base.py:119] [p0115] Starting concurrency...
  669. [0130 17:30:30 @base.py:198] Starting all threads & procs ...
  670. [0130 17:30:30 @base.py:122] [p0115] Setting default session
  671. [0130 17:30:30 @base.py:125] [p0115] Getting global step
  672. [0130 17:30:30 @base.py:127] [p0115] Start training with global_step=0
  673. [0130 17:30:31 @multigpu.py:323] [p0574]  step: count(21), step_time 1395.6, mean_step_time 1390.68, it/s 0.72
  674. [0130 17:30:31 @multigpu.py:323] [p0576]  step: count(21), step_time 1391.65, mean_step_time 1399.02, it/s 0.71
  675. sending to address tcp://p0112:61216
  676. ##### Sending to neptune:  online_score :  0.00865667495463 , 1.2 #####
  677. [u'online', 1.2]
  678. receiving
  679. [0130 17:30:33 @multigpu.py:323] [p0574]  step: count(22), step_time 1402.65, mean_step_time 1391.37, it/s 0.72
  680. [0130 17:30:33 @multigpu.py:323] [p0576]  step: count(22), step_time 1399.34, mean_step_time 1397.13, it/s 0.72
  681. [0130 17:30:33 @multigpu.py:323] [p0115]  step: count(1), step_time 42890.76, mean_step_time 42890.76, it/s 0.02
  682. [0130 17:30:34 @multigpu.py:323] [p0574]  step: count(23), step_time 1389.17, mean_step_time 1389.94, it/s 0.72
  683. [0130 17:30:34 @multigpu.py:323] [p0576]  step: count(23), step_time 1378.85, mean_step_time 1396.43, it/s 0.72
  684. [0130 17:30:34 @multigpu.py:323] [p0115]  step: count(2), step_time 1448.3, mean_step_time 22169.53, it/s 0.05
  685. [0130 17:30:35 @multigpu.py:323] [p0574]  step: count(24), step_time 1366.48, mean_step_time 1389.22, it/s 0.72
  686. [0130 17:30:35 @multigpu.py:323] [p0576]  step: count(24), step_time 1417.93, mean_step_time 1398.26, it/s 0.72
  687. [0130 17:30:36 @multigpu.py:323] [p0115]  step: count(3), step_time 1356.88, mean_step_time 15231.98, it/s 0.07
  688. [0130 17:30:37 @multigpu.py:323] [p0574]  step: count(25), step_time 1368.03, mean_step_time 1386.91, it/s 0.72
  689. [0130 17:30:37 @multigpu.py:323] [p0576]  step: count(25), step_time 1394.76, mean_step_time 1397.31, it/s 0.72
  690. [0130 17:30:37 @multigpu.py:323] [p0115]  step: count(4), step_time 1482.5, mean_step_time 11794.61, it/s 0.08
  691. [0130 17:30:38 @multigpu.py:323] [p0574]  step: count(26), step_time 1432.12, mean_step_time 1390.14, it/s 0.72
  692. [0130 17:30:38 @multigpu.py:323] [p0576]  step: count(26), step_time 1385.95, mean_step_time 1397.23, it/s 0.72
  693. [0130 17:30:38 @multigpu.py:323] [p0115]  step: count(5), step_time 1416.5, mean_step_time 9718.99, it/s 0.1
  694. [0130 17:30:40 @multigpu.py:323] [p0574]  step: count(27), step_time 1389.58, mean_step_time 1390.73, it/s 0.72
  695. [0130 17:30:40 @multigpu.py:323] [p0576]  step: count(27), step_time 1469.08, mean_step_time 1402.52, it/s 0.71
  696. [0130 17:30:40 @multigpu.py:323] [p0115]  step: count(6), step_time 1373.32, mean_step_time 8328.04, it/s 0.12
  697. [0130 17:30:41 @multigpu.py:323] [p0574]  step: count(28), step_time 1414.32, mean_step_time 1392.97, it/s 0.72
  698. [0130 17:30:41 @multigpu.py:323] [p0576]  step: count(28), step_time 1390.28, mean_step_time 1399.83, it/s 0.71
  699. [0130 17:30:41 @multigpu.py:323] [p0115]  step: count(7), step_time 1393.66, mean_step_time 7337.41, it/s 0.14
  700. [0130 17:30:42 @multigpu.py:323] [p0574]  step: count(29), step_time 1395.04, mean_step_time 1393.9, it/s 0.72
  701. sending to address tcp://p0112:61216
  702. ##### Sending to neptune:  online_score :  0.0116006455157 , 1.9 #####
  703. [u'online', 1.9]
  704. receiving
  705. [0130 17:30:43 @multigpu.py:323] [p0576]  step: count(29), step_time 1390.76, mean_step_time 1399.26, it/s 0.71
  706. [0130 17:30:43 @multigpu.py:323] [p0115]  step: count(8), step_time 1377.49, mean_step_time 6592.42, it/s 0.15
  707. [0130 17:30:44 @multigpu.py:323] [p0574]  step: count(30), step_time 1366.0, mean_step_time 1394.28, it/s 0.72
  708. [0130 17:30:44 @multigpu.py:323] [p0576]  step: count(30), step_time 1366.41, mean_step_time 1398.95, it/s 0.71
  709. [0130 17:30:44 @multigpu.py:323] [p0115]  step: count(9), step_time 1394.2, mean_step_time 6014.84, it/s 0.17
  710. [0130 17:30:45 @multigpu.py:323] [p0574]  step: count(31), step_time 1413.87, mean_step_time 1394.86, it/s 0.72
  711. [0130 17:30:45 @multigpu.py:323] [p0576]  step: count(31), step_time 1402.14, mean_step_time 1398.3, it/s 0.72
  712. [0130 17:30:45 @multigpu.py:323] [p0115]  step: count(10), step_time 1394.34, mean_step_time 5552.79, it/s 0.18
  713. [0130 17:30:47 @multigpu.py:323] [p0574]  step: count(32), step_time 1385.71, mean_step_time 1396.13, it/s 0.72
  714. [0130 17:30:47 @multigpu.py:323] [p0576]  step: count(32), step_time 1385.07, mean_step_time 1398.21, it/s 0.72
  715. [0130 17:30:47 @multigpu.py:323] [p0115]  step: count(11), step_time 1371.74, mean_step_time 5172.7, it/s 0.19
  716. [0130 17:30:48 @multigpu.py:323] [p0574]  step: count(33), step_time 1395.79, mean_step_time 1394.21, it/s 0.72
  717. [0130 17:30:48 @multigpu.py:323] [p0576]  step: count(33), step_time 1399.75, mean_step_time 1398.64, it/s 0.71
  718. [0130 17:30:48 @multigpu.py:323] [p0115]  step: count(12), step_time 1378.85, mean_step_time 4856.54, it/s 0.21
  719. [0130 17:30:49 @multigpu.py:323] [p0574]  step: count(34), step_time 1391.1, mean_step_time 1396.23, it/s 0.72
  720. [0130 17:30:49 @multigpu.py:323] [p0576]  step: count(34), step_time 1391.81, mean_step_time 1396.97, it/s 0.72
  721. [0130 17:30:50 @multigpu.py:323] [p0115]  step: count(13), step_time 1400.51, mean_step_time 4590.69, it/s 0.22
  722. [0130 17:30:51 @multigpu.py:323] [p0574]  step: count(35), step_time 1359.27, mean_step_time 1392.84, it/s 0.72
  723. [0130 17:30:51 @multigpu.py:323] [p0576]  step: count(35), step_time 1379.89, mean_step_time 1397.15, it/s 0.72
  724. [0130 17:30:51 @multigpu.py:323] [p0115]  step: count(14), step_time 1362.87, mean_step_time 4360.14, it/s 0.23
  725. [0130 17:30:52 @multigpu.py:323] [p0574]  step: count(36), step_time 1472.95, mean_step_time 1398.32, it/s 0.72
  726. [0130 17:30:52 @multigpu.py:323] [p0576]  step: count(36), step_time 1389.22, mean_step_time 1394.53, it/s 0.72
  727. [0130 17:30:52 @multigpu.py:323] [p0115]  step: count(15), step_time 1390.51, mean_step_time 4162.16, it/s 0.24
  728. [0130 17:30:54 @multigpu.py:323] [p0574]  step: count(37), step_time 1416.91, mean_step_time 1398.45, it/s 0.72
  729. [0130 17:30:54 @multigpu.py:323] [p0576]  step: count(37), step_time 1372.71, mean_step_time 1391.08, it/s 0.72
  730. [0130 17:30:54 @multigpu.py:323] [p0115]  step: count(16), step_time 1361.38, mean_step_time 3987.11, it/s 0.25
  731. [0130 17:30:55 @multigpu.py:323] [p0576]  step: count(38), step_time 1411.73, mean_step_time 1394.28, it/s 0.72
  732. [0130 17:30:55 @multigpu.py:323] [p0115]  step: count(17), step_time 1413.07, mean_step_time 3835.7, it/s 0.26
  733. [0130 17:30:55 @multigpu.py:323] [p0574]  step: count(38), step_time 1520.18, mean_step_time 1403.18, it/s 0.71
  734. [0130 17:30:56 @multigpu.py:323] [p0115]  step: count(18), step_time 1368.26, mean_step_time 3698.62, it/s 0.27
  735. [0130 17:30:56 @multigpu.py:323] [p0576]  step: count(39), step_time 1408.37, mean_step_time 1394.47, it/s 0.72
  736. [0130 17:30:57 @multigpu.py:323] [p0574]  step: count(39), step_time 1564.09, mean_step_time 1411.76, it/s 0.71
  737. [0130 17:30:58 @multigpu.py:323] [p0115]  step: count(19), step_time 1387.09, mean_step_time 3576.96, it/s 0.28
  738. [0130 17:30:58 @multigpu.py:323] [p0576]  step: count(40), step_time 1374.96, mean_step_time 1395.03, it/s 0.72
  739. [0130 17:30:58 @multigpu.py:323] [p0574]  step: count(40), step_time 1608.55, mean_step_time 1422.37, it/s 0.7
  740. [0130 17:30:59 @multigpu.py:323] [p0576]  step: count(41), step_time 1359.14, mean_step_time 1393.41, it/s 0.72
  741. [0130 17:30:59 @multigpu.py:323] [p0115]  step: count(20), step_time 1390.59, mean_step_time 3467.64, it/s 0.29
  742. [0130 17:31:00 @multigpu.py:323] [p0574]  step: count(41), step_time 1480.41, mean_step_time 1426.61, it/s 0.7
  743. [0130 17:31:01 @multigpu.py:323] [p0576]  step: count(42), step_time 1370.21, mean_step_time 1391.95, it/s 0.72
  744. [0130 17:31:01 @multigpu.py:323] [p0115]  step: count(21), step_time 1360.8, mean_step_time 1391.14, it/s 0.72
  745. [0130 17:31:01 @multigpu.py:323] [p0574]  step: count(42), step_time 1462.71, mean_step_time 1429.61, it/s 0.7
  746. [0130 17:31:02 @multigpu.py:323] [p0576]  step: count(43), step_time 1372.3, mean_step_time 1391.62, it/s 0.72
  747. [0130 17:31:02 @multigpu.py:323] [p0115]  step: count(22), step_time 1454.88, mean_step_time 1391.47, it/s 0.72
  748. sending to address tcp://p0112:61216
  749. ##### Sending to neptune:  online_score :  0.0171053697003 , 1.2 #####
  750. [u'online', 1.2]
  751. ##### Sending to neptune:  active_workers :  0.0171054766575 , 2 #####
  752. receiving
  753. [0130 17:31:03 @multigpu.py:323] [p0574]  step: count(43), step_time 1458.14, mean_step_time 1433.06, it/s 0.7
  754. [0130 17:31:03 @multigpu.py:323] [p0576]  step: count(44), step_time 1351.7, mean_step_time 1388.31, it/s 0.72
  755. [0130 17:31:03 @multigpu.py:323] [p0115]  step: count(23), step_time 1368.19, mean_step_time 1392.04, it/s 0.72
  756. [0130 17:31:04 @multigpu.py:323] [p0574]  step: count(44), step_time 1500.34, mean_step_time 1439.76, it/s 0.69
  757. [0130 17:31:05 @multigpu.py:323] [p0576]  step: count(45), step_time 1414.58, mean_step_time 1389.3, it/s 0.72
  758. [0130 17:31:05 @multigpu.py:323] [p0115]  step: count(24), step_time 1397.49, mean_step_time 1387.79, it/s 0.72
  759. [0130 17:31:06 @multigpu.py:323] [p0574]  step: count(45), step_time 1476.14, mean_step_time 1445.16, it/s 0.69
  760. [0130 17:31:06 @multigpu.py:323] [p0576]  step: count(46), step_time 1353.29, mean_step_time 1387.67, it/s 0.72
  761. [0130 17:31:06 @multigpu.py:323] [p0115]  step: count(25), step_time 1354.62, mean_step_time 1384.69, it/s 0.72
  762. [0130 17:31:07 @multigpu.py:323] [p0574]  step: count(46), step_time 1478.86, mean_step_time 1447.5, it/s 0.69
  763. [0130 17:31:07 @multigpu.py:323] [p0576]  step: count(47), step_time 1360.59, mean_step_time 1382.25, it/s 0.72
  764. [0130 17:31:08 @multigpu.py:323] [p0115]  step: count(26), step_time 1414.51, mean_step_time 1386.75, it/s 0.72
  765. [0130 17:31:09 @multigpu.py:323] [p0574]  step: count(47), step_time 1581.36, mean_step_time 1457.09, it/s 0.69
  766. [0130 17:31:10 @multigpu.py:323] [p0574]  step: count(48), step_time 1608.18, mean_step_time 1466.78, it/s 0.68
  767. [0130 17:31:10 @multigpu.py:323] [p0576]  step: count(48), step_time 2931.45, mean_step_time 1459.3, it/s 0.69
  768. [0130 17:31:10 @multigpu.py:323] [p0115]  step: count(27), step_time 2777.06, mean_step_time 1455.92, it/s 0.69
  769. sending to address tcp://p0112:61216
  770. ##### Sending to neptune:  online_score :  0.0195416252481 , 1.9 #####
  771. [u'online', 1.9]
  772. receiving
  773. [0130 17:31:12 @multigpu.py:323] [p0576]  step: count(49), step_time 1407.16, mean_step_time 1460.12, it/s 0.68
  774. [0130 17:31:12 @multigpu.py:323] [p0115]  step: count(28), step_time 1437.32, mean_step_time 1458.91, it/s 0.69
  775. [0130 17:31:12 @multigpu.py:323] [p0574]  step: count(49), step_time 1500.8, mean_step_time 1472.07, it/s 0.68
  776. [0130 17:31:13 @multigpu.py:323] [p0576]  step: count(50), step_time 1344.68, mean_step_time 1459.04, it/s 0.69
  777. [0130 17:31:13 @multigpu.py:323] [p0115]  step: count(29), step_time 1368.68, mean_step_time 1457.64, it/s 0.69
  778. [0130 17:31:13 @multigpu.py:323] [p0574]  step: count(50), step_time 1415.66, mean_step_time 1474.55, it/s 0.68
  779. [0130 17:31:14 @multigpu.py:323] [p0576]  step: count(51), step_time 1377.45, mean_step_time 1457.8, it/s 0.69
  780. [0130 17:31:14 @multigpu.py:323] [p0115]  step: count(30), step_time 1373.92, mean_step_time 1456.62, it/s 0.69
  781. [0130 17:31:15 @multigpu.py:323] [p0574]  step: count(51), step_time 1430.28, mean_step_time 1475.37, it/s 0.68
  782. sending to address tcp://p0112:61216
  783. ##### Sending to neptune:  online_score :  0.0207704544067 , 2.4 #####
  784. [u'online', 2.4]
  785. receiving
  786. [0130 17:31:16 @multigpu.py:323] [p0576]  step: count(52), step_time 1360.11, mean_step_time 1456.56, it/s 0.69
  787. [0130 17:31:16 @multigpu.py:323] [p0115]  step: count(31), step_time 1392.93, mean_step_time 1457.68, it/s 0.69
  788. [0130 17:31:16 @multigpu.py:323] [p0574]  step: count(52), step_time 1571.52, mean_step_time 1484.66, it/s 0.67
  789. [0130 17:31:17 @multigpu.py:323] [p0576]  step: count(53), step_time 1366.86, mean_step_time 1454.91, it/s 0.69
  790. [0130 17:31:17 @multigpu.py:323] [p0115]  step: count(32), step_time 1387.37, mean_step_time 1458.1, it/s 0.69
  791. [0130 17:31:18 @multigpu.py:323] [p0574]  step: count(53), step_time 1407.6, mean_step_time 1485.25, it/s 0.67
  792. [0130 17:31:19 @multigpu.py:323] [p0576]  step: count(54), step_time 1391.97, mean_step_time 1454.92, it/s 0.69
  793. [0130 17:31:19 @multigpu.py:323] [p0115]  step: count(33), step_time 1400.33, mean_step_time 1458.09, it/s 0.69
  794. [0130 17:31:19 @multigpu.py:323] [p0574]  step: count(54), step_time 1387.15, mean_step_time 1485.06, it/s 0.67
  795. [0130 17:31:20 @multigpu.py:323] [p0576]  step: count(55), step_time 1381.99, mean_step_time 1455.02, it/s 0.69
  796. [0130 17:31:20 @multigpu.py:323] [p0115]  step: count(34), step_time 1369.56, mean_step_time 1458.43, it/s 0.69
  797. [0130 17:31:20 @multigpu.py:323] [p0574]  step: count(55), step_time 1407.24, mean_step_time 1487.45, it/s 0.67
  798. [0130 17:31:21 @multigpu.py:323] [p0576]  step: count(56), step_time 1388.41, mean_step_time 1454.98, it/s 0.69
  799. [0130 17:31:21 @multigpu.py:323] [p0115]  step: count(35), step_time 1419.57, mean_step_time 1459.88, it/s 0.68
  800. [0130 17:31:22 @multigpu.py:323] [p0574]  step: count(56), step_time 1437.67, mean_step_time 1485.69, it/s 0.67
  801. [0130 17:31:23 @multigpu.py:323] [p0576]  step: count(57), step_time 1392.69, mean_step_time 1455.98, it/s 0.69
  802. [0130 17:31:23 @multigpu.py:323] [p0115]  step: count(36), step_time 1377.4, mean_step_time 1460.68, it/s 0.68
  803. [0130 17:31:23 @multigpu.py:323] [p0574]  step: count(57), step_time 1357.55, mean_step_time 1482.72, it/s 0.67
  804. [0130 17:31:24 @multigpu.py:323] [p0576]  step: count(58), step_time 1387.9, mean_step_time 1454.79, it/s 0.69
  805. [0130 17:31:24 @multigpu.py:323] [p0115]  step: count(37), step_time 1381.1, mean_step_time 1459.08, it/s 0.69
  806. sending to address tcp://p0112:61216
  807. ##### Sending to neptune:  online_score :  0.0233404641019 , 1.1 #####
  808. [u'online', 1.1]
  809. receiving
  810. [0130 17:31:25 @multigpu.py:323] [p0574]  step: count(58), step_time 1408.59, mean_step_time 1477.14, it/s 0.68
  811. [0130 17:31:26 @multigpu.py:323] [p0576]  step: count(59), step_time 1427.41, mean_step_time 1455.74, it/s 0.69
  812. [0130 17:31:26 @multigpu.py:323] [p0115]  step: count(38), step_time 1400.34, mean_step_time 1460.69, it/s 0.68
  813. [0130 17:31:26 @multigpu.py:323] [p0574]  step: count(59), step_time 1421.86, mean_step_time 1470.03, it/s 0.68
  814. [0130 17:31:27 @multigpu.py:323] [p0576]  step: count(60), step_time 1393.25, mean_step_time 1456.66, it/s 0.69
  815. [0130 17:31:27 @multigpu.py:323] [p0115]  step: count(39), step_time 1392.08, mean_step_time 1460.94, it/s 0.68
  816. [0130 17:31:27 @multigpu.py:323] [p0574]  step: count(60), step_time 1420.09, mean_step_time 1460.61, it/s 0.68
  817. [0130 17:31:28 @multigpu.py:323] [p0576]  step: count(61), step_time 1368.35, mean_step_time 1457.12, it/s 0.69
  818. [0130 17:31:28 @multigpu.py:323] [p0115]  step: count(40), step_time 1355.33, mean_step_time 1459.17, it/s 0.69
  819. [0130 17:31:29 @multigpu.py:323] [p0574]  step: count(61), step_time 1562.18, mean_step_time 1464.7, it/s 0.68
  820. [0130 17:31:30 @multigpu.py:323] [p0576]  step: count(62), step_time 1390.44, mean_step_time 1458.13, it/s 0.69
  821. [0130 17:31:30 @multigpu.py:323] [p0115]  step: count(41), step_time 1372.64, mean_step_time 1459.77, it/s 0.69
  822. [0130 17:31:31 @multigpu.py:323] [p0574]  step: count(62), step_time 1453.26, mean_step_time 1464.22, it/s 0.68
  823. [0130 17:31:31 @multigpu.py:323] [p0576]  step: count(63), step_time 1373.3, mean_step_time 1458.18, it/s 0.69
  824. [0130 17:31:31 @multigpu.py:323] [p0115]  step: count(42), step_time 1391.55, mean_step_time 1456.6, it/s 0.69
  825. [0130 17:31:32 @multigpu.py:323] [p0574]  step: count(63), step_time 1447.01, mean_step_time 1463.67, it/s 0.68
  826. [0130 17:31:32 @multigpu.py:323] [p0576]  step: count(64), step_time 1352.92, mean_step_time 1458.24, it/s 0.69
  827. [0130 17:31:32 @multigpu.py:323] [p0115]  step: count(43), step_time 1351.92, mean_step_time 1455.79, it/s 0.69
  828. [0130 17:31:34 @multigpu.py:323] [p0574]  step: count(64), step_time 1567.27, mean_step_time 1467.01, it/s 0.68
  829. [0130 17:31:34 @multigpu.py:323] [p0576]  step: count(65), step_time 1410.33, mean_step_time 1458.03, it/s 0.69
  830. [0130 17:31:34 @multigpu.py:323] [p0115]  step: count(44), step_time 1417.63, mean_step_time 1456.79, it/s 0.69
  831. [0130 17:31:35 @multigpu.py:323] [p0574]  step: count(65), step_time 1457.16, mean_step_time 1466.06, it/s 0.68
  832. [0130 17:31:35 @multigpu.py:323] [p0576]  step: count(66), step_time 1330.76, mean_step_time 1456.9, it/s 0.69
  833. [0130 17:31:35 @multigpu.py:323] [p0115]  step: count(45), step_time 1425.71, mean_step_time 1460.35, it/s 0.68
  834. sending to address tcp://p0112:61216
  835. ##### Sending to neptune:  online_score :  0.0265158143971 , 0.4 #####
  836. [u'online', 0.4]
  837. receiving
  838. [0130 17:31:36 @multigpu.py:323] [p0574]  step: count(66), step_time 1510.81, mean_step_time 1467.66, it/s 0.68
  839. [0130 17:31:37 @multigpu.py:323] [p0576]  step: count(67), step_time 1398.99, mean_step_time 1458.82, it/s 0.69
  840. [0130 17:31:37 @multigpu.py:323] [p0115]  step: count(46), step_time 1376.59, mean_step_time 1458.45, it/s 0.69
  841. [0130 17:31:38 @multigpu.py:323] [p0576]  step: count(68), step_time 1353.56, mean_step_time 1379.93, it/s 0.72
  842. [0130 17:31:39 @multigpu.py:323] [p0574]  step: count(67), step_time 2849.34, mean_step_time 1531.06, it/s 0.65
  843. [0130 17:31:39 @multigpu.py:323] [p0576]  step: count(69), step_time 1413.17, mean_step_time 1380.23, it/s 0.72
  844. [0130 17:31:39 @multigpu.py:323] [p0115]  step: count(47), step_time 2621.15, mean_step_time 1450.66, it/s 0.69
  845. [0130 17:31:41 @multigpu.py:323] [p0576]  step: count(70), step_time 1364.24, mean_step_time 1381.21, it/s 0.72
  846. [0130 17:31:41 @multigpu.py:323] [p0574]  step: count(68), step_time 1413.99, mean_step_time 1521.35, it/s 0.66
  847. [0130 17:31:41 @multigpu.py:323] [p0115]  step: count(48), step_time 1445.55, mean_step_time 1451.07, it/s 0.69
  848. sending to address tcp://p0112:61216
  849. ##### Sending to neptune:  online_score :  0.0280429810948 , 1.2 #####
  850. [u'online', 1.2]
  851. receiving
  852. [0130 17:31:42 @multigpu.py:323] [p0576]  step: count(71), step_time 1389.69, mean_step_time 1381.82, it/s 0.72
  853. [0130 17:31:42 @multigpu.py:323] [p0574]  step: count(69), step_time 1360.72, mean_step_time 1514.35, it/s 0.66
  854. [0130 17:31:42 @multigpu.py:323] [p0115]  step: count(49), step_time 1398.37, mean_step_time 1452.55, it/s 0.69
  855. [0130 17:31:43 @multigpu.py:323] [p0576]  step: count(72), step_time 1337.14, mean_step_time 1380.67, it/s 0.72
  856. [0130 17:31:44 @multigpu.py:323] [p0574]  step: count(70), step_time 1395.54, mean_step_time 1513.34, it/s 0.66
  857. [0130 17:31:44 @multigpu.py:323] [p0115]  step: count(50), step_time 1376.98, mean_step_time 1452.7, it/s 0.69
  858. [0130 17:31:45 @multigpu.py:323] [p0576]  step: count(73), step_time 1400.4, mean_step_time 1382.35, it/s 0.72
  859. [0130 17:31:45 @multigpu.py:323] [p0574]  step: count(71), step_time 1398.99, mean_step_time 1511.78, it/s 0.66
  860. [0130 17:31:45 @multigpu.py:323] [p0115]  step: count(51), step_time 1356.44, mean_step_time 1450.88, it/s 0.69
  861. [0130 17:31:46 @multigpu.py:323] [p0576]  step: count(74), step_time 1366.03, mean_step_time 1381.05, it/s 0.72
  862. [0130 17:31:46 @multigpu.py:323] [p0574]  step: count(72), step_time 1349.11, mean_step_time 1500.66, it/s 0.67
  863. [0130 17:31:46 @multigpu.py:323] [p0115]  step: count(52), step_time 1407.75, mean_step_time 1451.9, it/s 0.69
  864. [0130 17:31:48 @multigpu.py:323] [p0574]  step: count(73), step_time 1410.12, mean_step_time 1500.78, it/s 0.67
  865. [0130 17:31:48 @multigpu.py:323] [p0576]  step: count(75), step_time 1482.05, mean_step_time 1386.05, it/s 0.72
  866. [0130 17:31:48 @multigpu.py:323] [p0115]  step: count(53), step_time 1391.0, mean_step_time 1451.43, it/s 0.69
  867. [0130 17:31:49 @multigpu.py:323] [p0115]  step: count(54), step_time 1359.58, mean_step_time 1450.93, it/s 0.69
  868. [0130 17:31:49 @multigpu.py:323] [p0576]  step: count(76), step_time 1401.94, mean_step_time 1386.73, it/s 0.72
  869. [0130 17:31:49 @multigpu.py:323] [p0574]  step: count(74), step_time 1418.6, mean_step_time 1502.36, it/s 0.67
  870. sending to address tcp://p0112:61216
  871. ##### Sending to neptune:  online_score :  0.0303178360727 , 1.0 #####
  872. [u'online', 1.0]
  873. receiving
  874. [0130 17:31:50 @multigpu.py:323] [p0576]  step: count(77), step_time 1349.13, mean_step_time 1384.55, it/s 0.72
  875. [0130 17:31:50 @multigpu.py:323] [p0574]  step: count(75), step_time 1366.17, mean_step_time 1500.3, it/s 0.67
  876. [0130 17:31:50 @multigpu.py:323] [p0115]  step: count(55), step_time 1408.07, mean_step_time 1450.36, it/s 0.69
  877. [0130 17:31:52 @multigpu.py:323] [p0574]  step: count(76), step_time 1361.83, mean_step_time 1496.51, it/s 0.67
  878. [0130 17:31:52 @multigpu.py:323] [p0576]  step: count(78), step_time 1419.72, mean_step_time 1386.14, it/s 0.72
  879. [0130 17:31:52 @multigpu.py:323] [p0115]  step: count(56), step_time 1383.6, mean_step_time 1450.67, it/s 0.69
  880. [0130 17:31:53 @multigpu.py:323] [p0574]  step: count(77), step_time 1408.0, mean_step_time 1499.03, it/s 0.67
  881. [0130 17:31:53 @multigpu.py:323] [p0576]  step: count(79), step_time 1385.7, mean_step_time 1384.06, it/s 0.72
  882. [0130 17:31:53 @multigpu.py:323] [p0115]  step: count(57), step_time 1368.43, mean_step_time 1450.04, it/s 0.69
  883. [0130 17:31:55 @multigpu.py:323] [p0115]  step: count(58), step_time 1342.06, mean_step_time 1447.12, it/s 0.69
  884. [0130 17:31:55 @multigpu.py:323] [p0574]  step: count(78), step_time 1385.83, mean_step_time 1497.89, it/s 0.67
  885. [0130 17:31:55 @multigpu.py:323] [p0576]  step: count(80), step_time 1390.21, mean_step_time 1383.9, it/s 0.72
  886. [0130 17:31:56 @multigpu.py:323] [p0115]  step: count(59), step_time 1394.26, mean_step_time 1447.23, it/s 0.69
  887. [0130 17:31:56 @multigpu.py:323] [p0576]  step: count(81), step_time 1366.46, mean_step_time 1383.81, it/s 0.72
  888. [0130 17:31:56 @multigpu.py:323] [p0574]  step: count(79), step_time 1432.96, mean_step_time 1498.45, it/s 0.67
  889. sending to address tcp://p0112:61216
  890. ##### Sending to neptune:  online_score :  0.0322564772103 , 0.7 #####
  891. [u'online', 0.7]
  892. receiving
  893. [0130 17:31:57 @multigpu.py:323] [p0576]  step: count(82), step_time 1393.91, mean_step_time 1383.98, it/s 0.72
  894. [0130 17:31:57 @multigpu.py:323] [p0115]  step: count(60), step_time 1417.24, mean_step_time 1450.33, it/s 0.69
  895. [0130 17:31:57 @multigpu.py:323] [p0574]  step: count(80), step_time 1353.39, mean_step_time 1495.11, it/s 0.67
  896. [0130 17:31:59 @multigpu.py:323] [p0576]  step: count(83), step_time 1352.62, mean_step_time 1382.95, it/s 0.72
  897. [0130 17:31:59 @multigpu.py:323] [p0115]  step: count(61), step_time 1369.22, mean_step_time 1450.16, it/s 0.69
  898. [0130 17:31:59 @multigpu.py:323] [p0574]  step: count(81), step_time 1419.05, mean_step_time 1487.96, it/s 0.67
  899. [0130 17:32:00 @multigpu.py:323] [p0576]  step: count(84), step_time 1376.94, mean_step_time 1384.15, it/s 0.72
  900. [0130 17:32:00 @multigpu.py:323] [p0115]  step: count(62), step_time 1392.11, mean_step_time 1450.18, it/s 0.69
  901. [0130 17:32:00 @multigpu.py:323] [p0574]  step: count(82), step_time 1567.51, mean_step_time 1493.67, it/s 0.67
  902. [0130 17:32:01 @multigpu.py:323] [p0576]  step: count(85), step_time 1364.35, mean_step_time 1381.85, it/s 0.72
  903. [0130 17:32:02 @multigpu.py:323] [p0115]  step: count(63), step_time 1402.84, mean_step_time 1452.73, it/s 0.69
  904. [0130 17:32:02 @multigpu.py:323] [p0574]  step: count(83), step_time 1388.35, mean_step_time 1490.74, it/s 0.67
  905. [0130 17:32:03 @multigpu.py:323] [p0576]  step: count(86), step_time 1388.53, mean_step_time 1384.74, it/s 0.72
  906. [0130 17:32:03 @multigpu.py:323] [p0115]  step: count(64), step_time 1359.18, mean_step_time 1449.81, it/s 0.69
  907. [0130 17:32:03 @multigpu.py:323] [p0574]  step: count(84), step_time 1357.99, mean_step_time 1480.27, it/s 0.68
  908. [0130 17:32:04 @multigpu.py:323] [p0576]  step: count(87), step_time 1357.1, mean_step_time 1382.64, it/s 0.72
  909. [0130 17:32:04 @multigpu.py:323] [p0115]  step: count(65), step_time 1433.11, mean_step_time 1450.18, it/s 0.69
  910. [0130 17:32:05 @multigpu.py:323] [p0574]  step: count(85), step_time 1411.06, mean_step_time 1477.97, it/s 0.68
  911. [0130 17:32:06 @multigpu.py:323] [p0576]  step: count(88), step_time 1413.38, mean_step_time 1385.63, it/s 0.72
  912. [0130 17:32:06 @multigpu.py:323] [p0115]  step: count(66), step_time 1376.56, mean_step_time 1450.18, it/s 0.69
  913. [0130 17:32:06 @multigpu.py:323] [p0574]  step: count(86), step_time 1371.07, mean_step_time 1470.98, it/s 0.68
  914. sending to address tcp://p0112:61216
  915. ##### Sending to neptune:  online_score :  0.0349453416136 , 1.0 #####
  916. [u'online', 1.0]
  917. ##### Sending to neptune:  active_workers :  0.034945417775 , 3 #####
  918. receiving
  919. [0130 17:32:07 @multigpu.py:323] [p0574]  step: count(87), step_time 1381.74, mean_step_time 1397.6, it/s 0.72
  920. [0130 17:32:07 @multigpu.py:323] [p0576]  step: count(89), step_time 1651.44, mean_step_time 1397.55, it/s 0.72
  921. [0130 17:32:07 @multigpu.py:323] [p0115]  step: count(67), step_time 1568.35, mean_step_time 1397.54, it/s 0.72
  922. [0130 17:32:09 @multigpu.py:323] [p0574]  step: count(88), step_time 1388.4, mean_step_time 1396.32, it/s 0.72
  923. [0130 17:32:09 @multigpu.py:323] [p0115]  step: count(68), step_time 1399.14, mean_step_time 1395.21, it/s 0.72
  924. [0130 17:32:09 @multigpu.py:323] [p0576]  step: count(90), step_time 1415.9, mean_step_time 1400.13, it/s 0.71
  925. [0130 17:32:10 @multigpu.py:323] [p0574]  step: count(89), step_time 1346.75, mean_step_time 1395.62, it/s 0.72
  926. [0130 17:32:10 @multigpu.py:323] [p0576]  step: count(91), step_time 1336.87, mean_step_time 1397.49, it/s 0.72
  927. [0130 17:32:10 @multigpu.py:323] [p0115]  step: count(69), step_time 1472.22, mean_step_time 1398.91, it/s 0.71
  928. [0130 17:32:11 @multigpu.py:323] [p0574]  step: count(90), step_time 1396.78, mean_step_time 1395.68, it/s 0.72
  929. [0130 17:32:12 @multigpu.py:323] [p0576]  step: count(92), step_time 1474.83, mean_step_time 1404.37, it/s 0.71
  930. [0130 17:32:12 @multigpu.py:323] [p0115]  step: count(70), step_time 1359.43, mean_step_time 1398.03, it/s 0.72
  931. [0130 17:32:13 @multigpu.py:323] [p0574]  step: count(91), step_time 1389.08, mean_step_time 1395.19, it/s 0.72
  932. [0130 17:32:13 @multigpu.py:323] [p0576]  step: count(93), step_time 1346.17, mean_step_time 1401.66, it/s 0.71
  933. [0130 17:32:13 @multigpu.py:323] [p0115]  step: count(71), step_time 1420.7, mean_step_time 1401.24, it/s 0.71
  934. [0130 17:32:14 @multigpu.py:323] [p0574]  step: count(92), step_time 1377.89, mean_step_time 1396.63, it/s 0.72
  935. [0130 17:32:14 @multigpu.py:323] [p0576]  step: count(94), step_time 1409.36, mean_step_time 1403.83, it/s 0.71
  936. [0130 17:32:14 @multigpu.py:323] [p0115]  step: count(72), step_time 1387.85, mean_step_time 1400.25, it/s 0.71
  937. sending to address tcp://p0112:61216
  938. ##### Sending to neptune:  online_score :  0.0373082438442 , 1.3 #####
  939. [u'online', 1.3]
  940. receiving
  941. [0130 17:32:16 @multigpu.py:323] [p0574]  step: count(93), step_time 1398.81, mean_step_time 1396.06, it/s 0.72
  942. [0130 17:32:16 @multigpu.py:323] [p0576]  step: count(95), step_time 1378.33, mean_step_time 1398.64, it/s 0.71
  943. [0130 17:32:16 @multigpu.py:323] [p0115]  step: count(73), step_time 1430.54, mean_step_time 1402.23, it/s 0.71
  944. [0130 17:32:17 @multigpu.py:323] [p0576]  step: count(96), step_time 1360.46, mean_step_time 1396.57, it/s 0.72
  945. [0130 17:32:17 @multigpu.py:323] [p0574]  step: count(94), step_time 1466.82, mean_step_time 1398.47, it/s 0.72
  946. [0130 17:32:17 @multigpu.py:323] [p0115]  step: count(74), step_time 1409.39, mean_step_time 1404.72, it/s 0.71
  947. [0130 17:32:18 @multigpu.py:323] [p0576]  step: count(97), step_time 1391.39, mean_step_time 1398.68, it/s 0.71
  948. [0130 17:32:19 @multigpu.py:323] [p0574]  step: count(95), step_time 1440.57, mean_step_time 1402.19, it/s 0.71
  949. [0130 17:32:19 @multigpu.py:323] [p0115]  step: count(75), step_time 1384.89, mean_step_time 1403.56, it/s 0.71
  950. [0130 17:32:20 @multigpu.py:323] [p0576]  step: count(98), step_time 1356.32, mean_step_time 1395.51, it/s 0.72
  951. [0130 17:32:20 @multigpu.py:323] [p0115]  step: count(76), step_time 1322.07, mean_step_time 1400.48, it/s 0.71
  952. [0130 17:32:20 @multigpu.py:323] [p0574]  step: count(96), step_time 1402.93, mean_step_time 1404.25, it/s 0.71
  953. [0130 17:32:21 @multigpu.py:323] [p0576]  step: count(99), step_time 1364.49, mean_step_time 1394.45, it/s 0.72
  954. [0130 17:32:21 @multigpu.py:323] [p0115]  step: count(77), step_time 1393.59, mean_step_time 1401.74, it/s 0.71
  955. [0130 17:32:21 @multigpu.py:323] [p0574]  step: count(97), step_time 1396.28, mean_step_time 1403.66, it/s 0.71
  956. [0130 17:32:23 @multigpu.py:323] [p0576]  step: count(100), step_time 1392.71, mean_step_time 1394.58, it/s 0.72
  957. sending debugging info...
  958. sending to address tcp://p0112:61216
  959. ##### Sending to neptune:  mean_delay :  0.0394246213304 , 0.0 #####
  960. sending to address tcp://p0112:61216
  961. ##### Sending to neptune:  max_delay :  0.0394246213304 , -0.0 #####
  962. ##### Sending to neptune:  min_delay :  0.0394246213304 , -0.0 #####
  963. [u'delays', [0.0, -0.0, -0.0]]
  964. receiving
  965. ##### Sending to neptune:  cost :  0.0394251494275 , -0.0056879054755 #####
  966. sending to address tcp://p0112:61216
  967. ##### Sending to neptune:  policy_loss :  0.0394251494275 , 0.399972140789 #####
  968. ##### Sending to neptune:  xentropy_loss :  0.0394251494275 , -2.29181861877 #####
  969. ##### Sending to neptune:  value_loss :  0.0394251494275 , 1.16379439831 #####
  970. ##### Sending to neptune:  advantage :  0.0394251494275 , -0.00176788156386 #####
  971. ##### Sending to neptune:  pred_reward :  0.0394251494275 , 0.261878997087 #####
  972. ##### Sending to neptune:  max_logit :  0.0394251494275 , 0.181682661176 #####
  973. [u'loss', -0.005687905475497246, 0.399972140789032, -2.291818618774414, 1.1637943983078003, -0.0017678815638646483, 0.26187899708747864, 0.18168266117572784]
  974. receiving
  975. ##### Sending to neptune:  active_relus :  0.0394261499908 , 9003946.08 #####
  976. ##### Sending to neptune:  dp_per_s :  0.0394261499908 , 86.5037380059 #####
  977. [u'other', 9003946.08, 86.50373800587406]
  978. receiving
  979. [0130 17:32:23 @multigpu.py:323] [p0115]  step: count(78), step_time 1362.33, mean_step_time 1402.75, it/s 0.71
  980. [0130 17:32:23 @multigpu.py:323] [p0574]  step: count(98), step_time 1373.19, mean_step_time 1403.03, it/s 0.71
  981. [0130 17:32:24 @multigpu.py:323] [p0576]  step: count(101), step_time 1420.15, mean_step_time 1397.26, it/s 0.72
  982. [0130 17:32:24 @multigpu.py:323] [p0115]  step: count(79), step_time 1413.0, mean_step_time 1403.69, it/s 0.71
  983. [0130 17:32:24 @multigpu.py:323] [p0574]  step: count(99), step_time 1382.8, mean_step_time 1400.52, it/s 0.71
  984. sending to address tcp://p0112:61216
  985. ##### Sending to neptune:  online_score :  0.040045074953 , 0.8 #####
  986. [u'online', 0.8]
  987. receiving
  988. [0130 17:32:25 @multigpu.py:323] [p0576]  step: count(102), step_time 1406.42, mean_step_time 1397.89, it/s 0.72
  989. [0130 17:32:25 @multigpu.py:323] [p0115]  step: count(80), step_time 1359.02, mean_step_time 1400.78, it/s 0.71
  990. [0130 17:32:25 @multigpu.py:323] [p0574]  step: count(100), step_time 1389.18, mean_step_time 1402.31, it/s 0.71
  991. sending debugging info...
  992. sending to address tcp://p0112:61216
  993. ##### Sending to neptune:  mean_delay :  0.0402365121577 , 0.0 #####
  994. ##### Sending to neptune:  max_delay :  0.0402365121577 , -0.0 #####
  995. sending to address tcp://p0112:61216
  996. ##### Sending to neptune:  min_delay :  0.0402365121577 , -0.0 #####
  997. [u'delays', [0.0, -0.0, -0.0]]
  998. receiving
  999. ##### Sending to neptune:  cost :  0.0402370121744 , 0.00216898717918 #####
  1000. ##### Sending to neptune:  policy_loss :  0.0402370121744 , 1.23316562176 #####
  1001. sending to address tcp://p0112:61216
  1002. ##### Sending to neptune:  xentropy_loss :  0.0402370121744 , -2.29180669785 #####
  1003. ##### Sending to neptune:  value_loss :  0.0402370121744 , 1.3362711668 #####
  1004. ##### Sending to neptune:  advantage :  0.0402370121744 , -0.00537997484207 #####
  1005. ##### Sending to neptune:  pred_reward :  0.0402370121744 , 0.260432839394 #####
  1006. ##### Sending to neptune:  max_logit :  0.0402370121744 , 0.181708931923 #####
  1007. [u'loss', 0.0021689871791750193, 1.2331656217575073, -2.291806697845459, 1.3362711668014526, -0.005379974842071533, 0.2604328393936157, 0.1817089319229126]
  1008. receiving
  1009. ##### Sending to neptune:  active_relus :  0.0402375141117 , 8991974.32 #####
  1010. ##### Sending to neptune:  dp_per_s :  0.0402375141117 , 84.6942909111 #####
  1011. [u'other', 8991974.32, 84.69429091109998]
  1012. receiving
  1013. [0130 17:32:27 @multigpu.py:323] [p0576]  step: count(103), step_time 1386.27, mean_step_time 1399.57, it/s 0.71
  1014. [0130 17:32:27 @multigpu.py:323] [p0115]  step: count(81), step_time 1420.97, mean_step_time 1403.37, it/s 0.71
  1015. [0130 17:32:27 @multigpu.py:323] [p0574]  step: count(101), step_time 1417.69, mean_step_time 1402.24, it/s 0.71
  1016. [0130 17:32:28 @multigpu.py:323] [p0576]  step: count(104), step_time 1378.61, mean_step_time 1399.65, it/s 0.71
  1017. [0130 17:32:28 @multigpu.py:323] [p0574]  step: count(102), step_time 1342.28, mean_step_time 1390.98, it/s 0.72
  1018. [0130 17:32:28 @multigpu.py:323] [p0115]  step: count(82), step_time 1416.15, mean_step_time 1404.57, it/s 0.71
  1019. [0130 17:32:30 @multigpu.py:323] [p0576]  step: count(105), step_time 1376.33, mean_step_time 1400.25, it/s 0.71
  1020. [0130 17:32:30 @multigpu.py:323] [p0574]  step: count(103), step_time 1393.23, mean_step_time 1391.23, it/s 0.72
  1021. [0130 17:32:30 @multigpu.py:323] [p0115]  step: count(83), step_time 1367.98, mean_step_time 1402.82, it/s 0.71
  1022. [0130 17:32:31 @multigpu.py:323] [p0576]  step: count(106), step_time 1405.28, mean_step_time 1401.09, it/s 0.71
  1023. [0130 17:32:31 @multigpu.py:323] [p0574]  step: count(104), step_time 1371.23, mean_step_time 1391.89, it/s 0.72
  1024. [0130 17:32:31 @multigpu.py:323] [p0115]  step: count(84), step_time 1419.1, mean_step_time 1405.82, it/s 0.71
  1025. [0130 17:32:32 @multigpu.py:323] [p0576]  step: count(107), step_time 1391.55, mean_step_time 1402.81, it/s 0.71
  1026. [0130 17:32:32 @multigpu.py:323] [p0574]  step: count(105), step_time 1418.55, mean_step_time 1392.26, it/s 0.72
  1027. [0130 17:32:32 @multigpu.py:323] [p0115]  step: count(85), step_time 1394.13, mean_step_time 1403.87, it/s 0.71
  1028. [0130 17:32:34 @multigpu.py:323] [p0576]  step: count(108), step_time 1447.96, mean_step_time 1404.54, it/s 0.71
  1029. [0130 17:32:34 @multigpu.py:323] [p0574]  step: count(106), step_time 1361.46, mean_step_time 1391.78, it/s 0.72
  1030. [0130 17:32:34 @multigpu.py:323] [p0115]  step: count(86), step_time 1395.83, mean_step_time 1404.84, it/s 0.71
  1031. sending to address tcp://p0112:61216
  1032. ##### Sending to neptune:  online_score :  0.0427628719144 , 2.1 #####
  1033. [u'online', 2.1]
  1034. receiving
  1035. [0130 17:32:35 @multigpu.py:323] [p0574]  step: count(107), step_time 1429.53, mean_step_time 1394.17, it/s 0.72
  1036. [0130 17:32:35 @multigpu.py:323] [p0576]  step: count(109), step_time 1442.03, mean_step_time 1394.07, it/s 0.72
  1037. [0130 17:32:35 @multigpu.py:323] [p0115]  step: count(87), step_time 1353.86, mean_step_time 1394.11, it/s 0.72
  1038. [0130 17:32:37 @multigpu.py:323] [p0576]  step: count(110), step_time 1348.97, mean_step_time 1390.72, it/s 0.72
  1039. [0130 17:32:37 @multigpu.py:323] [p0574]  step: count(108), step_time 1386.77, mean_step_time 1394.09, it/s 0.72
  1040. [0130 17:32:37 @multigpu.py:323] [p0115]  step: count(88), step_time 1412.35, mean_step_time 1394.77, it/s 0.72
  1041. [0130 17:32:38 @multigpu.py:323] [p0576]  step: count(111), step_time 1377.92, mean_step_time 1392.78, it/s 0.72
  1042. [0130 17:32:38 @multigpu.py:323] [p0115]  step: count(89), step_time 1370.59, mean_step_time 1389.69, it/s 0.72
  1043. [0130 17:32:38 @multigpu.py:323] [p0574]  step: count(109), step_time 1411.26, mean_step_time 1397.32, it/s 0.72
  1044. [0130 17:32:39 @multigpu.py:323] [p0576]  step: count(112), step_time 1386.66, mean_step_time 1388.37, it/s 0.72
  1045. [0130 17:32:39 @multigpu.py:323] [p0115]  step: count(90), step_time 1351.87, mean_step_time 1389.31, it/s 0.72
  1046. [0130 17:32:39 @multigpu.py:323] [p0574]  step: count(110), step_time 1452.75, mean_step_time 1400.11, it/s 0.71
  1047. [0130 17:32:41 @multigpu.py:323] [p0115]  step: count(91), step_time 1359.65, mean_step_time 1386.26, it/s 0.72
  1048. [0130 17:32:41 @multigpu.py:323] [p0576]  step: count(113), step_time 1427.16, mean_step_time 1392.42, it/s 0.72
  1049. [0130 17:32:41 @multigpu.py:323] [p0574]  step: count(111), step_time 1353.86, mean_step_time 1398.35, it/s 0.72
  1050. sending to address tcp://p0112:61216
  1051. ##### Sending to neptune:  online_score :  0.0445858110984 , 1.5 #####
  1052. [u'online', 1.5]
  1053. receiving
  1054. [0130 17:32:42 @multigpu.py:323] [p0576]  step: count(114), step_time 1351.34, mean_step_time 1389.52, it/s 0.72
  1055. [0130 17:32:42 @multigpu.py:323] [p0115]  step: count(92), step_time 1456.38, mean_step_time 1389.68, it/s 0.72
  1056. [0130 17:32:42 @multigpu.py:323] [p0574]  step: count(112), step_time 1408.49, mean_step_time 1399.88, it/s 0.71
  1057. [0130 17:32:44 @multigpu.py:323] [p0115]  step: count(93), step_time 1363.97, mean_step_time 1386.36, it/s 0.72
  1058. [0130 17:32:44 @multigpu.py:323] [p0576]  step: count(115), step_time 1460.08, mean_step_time 1393.61, it/s 0.72
  1059. [0130 17:32:44 @multigpu.py:323] [p0574]  step: count(113), step_time 1432.72, mean_step_time 1401.58, it/s 0.71
  1060. [0130 17:32:45 @multigpu.py:323] [p0115]  step: count(94), step_time 1370.03, mean_step_time 1384.39, it/s 0.72
  1061. [0130 17:32:45 @multigpu.py:323] [p0576]  step: count(116), step_time 1421.28, mean_step_time 1396.65, it/s 0.72
  1062. [0130 17:32:45 @multigpu.py:323] [p0574]  step: count(114), step_time 1433.08, mean_step_time 1399.89, it/s 0.71
  1063. [0130 17:32:46 @multigpu.py:323] [p0115]  step: count(95), step_time 1378.92, mean_step_time 1384.09, it/s 0.72
  1064. [0130 17:32:46 @multigpu.py:323] [p0576]  step: count(117), step_time 1382.6, mean_step_time 1396.21, it/s 0.72
  1065. [0130 17:32:47 @multigpu.py:323] [p0574]  step: count(115), step_time 1430.45, mean_step_time 1399.39, it/s 0.71
  1066. [0130 17:32:48 @multigpu.py:323] [p0115]  step: count(96), step_time 1401.86, mean_step_time 1388.08, it/s 0.72
  1067. [0130 17:32:48 @multigpu.py:323] [p0576]  step: count(118), step_time 1369.23, mean_step_time 1396.85, it/s 0.72
  1068. [0130 17:32:48 @multigpu.py:323] [p0574]  step: count(116), step_time 1421.34, mean_step_time 1400.31, it/s 0.71
  1069. [0130 17:32:49 @multigpu.py:323] [p0115]  step: count(97), step_time 1382.13, mean_step_time 1387.51, it/s 0.72
  1070. [0130 17:32:49 @multigpu.py:323] [p0576]  step: count(119), step_time 1389.47, mean_step_time 1398.1, it/s 0.72
  1071. [0130 17:32:49 @multigpu.py:323] [p0574]  step: count(117), step_time 1401.27, mean_step_time 1400.56, it/s 0.71
  1072. [0130 17:32:50 @multigpu.py:323] [p0115]  step: count(98), step_time 1443.72, mean_step_time 1391.58, it/s 0.72
  1073. [0130 17:32:51 @multigpu.py:323] [p0576]  step: count(120), step_time 1429.82, mean_step_time 1399.96, it/s 0.71
  1074. [0130 17:32:51 @multigpu.py:323] [p0574]  step: count(118), step_time 1454.38, mean_step_time 1404.62, it/s 0.71
  1075. [0130 17:32:52 @multigpu.py:323] [p0115]  step: count(99), step_time 1387.8, mean_step_time 1390.31, it/s 0.72
  1076. [0130 17:32:52 @multigpu.py:323] [p0576]  step: count(121), step_time 1392.97, mean_step_time 1398.6, it/s 0.72
  1077. [0130 17:32:52 @multigpu.py:323] [p0574]  step: count(119), step_time 1527.24, mean_step_time 1411.84, it/s 0.71
  1078. [0130 17:32:53 @multigpu.py:323] [p0115]  step: count(100), step_time 1381.26, mean_step_time 1391.43, it/s 0.72
  1079. sending debugging info...
  1080. sending to address tcp://p0112:61216
  1081. ##### Sending to neptune:  mean_delay :  0.047960005535 , 0.0 #####
  1082. ##### Sending to neptune:  max_delay :  0.047960005535 , -0.0 #####
  1083. sending to address tcp://p0112:61216
  1084. ##### Sending to neptune:  min_delay :  0.047960005535 , -0.0 #####
  1085. [u'delays', [0.0, -0.0, -0.0]]
  1086. receiving
  1087. ##### Sending to neptune:  cost :  0.0479606341653 , -0.0105593772605 #####
  1088. sending to address tcp://p0112:61216
  1089. ##### Sending to neptune:  policy_loss :  0.0479606341653 , -0.167208120227 #####
  1090. ##### Sending to neptune:  xentropy_loss :  0.0479606341653 , -2.29150485992 #####
  1091. ##### Sending to neptune:  value_loss :  0.0479606341653 , 1.10711240768 #####
  1092. ##### Sending to neptune:  advantage :  0.0479606341653 , 0.000706825871021 #####
  1093. ##### Sending to neptune:  pred_reward :  0.0479606341653 , 0.30842140317 #####
  1094. ##### Sending to neptune:  max_logit :  0.0479606341653 , 0.18301834166 #####
  1095. [u'loss', -0.010559377260506153, -0.16720812022686005, -2.2915048599243164, 1.1071124076843262, 0.0007068258710205555, 0.30842140316963196, 0.18301834166049957]
  1096. receiving
  1097. ##### Sending to neptune:  active_relus :  0.0479611127244 , 8962460.69 #####
  1098. ##### Sending to neptune:  dp_per_s :  0.0479611127244 , 76.3888871862 #####
  1099. [u'other', 8962460.69, 76.38888718623758]
  1100. receiving
  1101. [0130 17:32:53 @multigpu.py:323] [p0576]  step: count(122), step_time 1392.28, mean_step_time 1397.89, it/s 0.72
  1102. [0130 17:32:54 @multigpu.py:323] [p0574]  step: count(120), step_time 1530.71, mean_step_time 1418.91, it/s 0.7
  1103. sending to address tcp://p0112:61216
  1104. ##### Sending to neptune:  online_score :  0.0481446746985 , 2.2 #####
  1105. [u'online', 2.2]
  1106. receiving
  1107. [0130 17:32:55 @multigpu.py:323] [p0115]  step: count(101), step_time 1410.22, mean_step_time 1390.89, it/s 0.72
  1108. [0130 17:32:55 @multigpu.py:323] [p0576]  step: count(123), step_time 1348.5, mean_step_time 1396.0, it/s 0.72
  1109. [0130 17:32:55 @multigpu.py:323] [p0574]  step: count(121), step_time 1452.44, mean_step_time 1420.65, it/s 0.7
  1110. [0130 17:32:56 @multigpu.py:323] [p0115]  step: count(102), step_time 1357.53, mean_step_time 1387.96, it/s 0.72
  1111. [0130 17:32:56 @multigpu.py:323] [p0576]  step: count(124), step_time 1363.25, mean_step_time 1395.23, it/s 0.72
  1112. [0130 17:32:57 @multigpu.py:323] [p0574]  step: count(122), step_time 1581.99, mean_step_time 1432.64, it/s 0.7
  1113. [0130 17:32:57 @multigpu.py:323] [p0115]  step: count(103), step_time 1383.49, mean_step_time 1388.73, it/s 0.72
  1114. [0130 17:32:57 @multigpu.py:323] [p0576]  step: count(125), step_time 1443.28, mean_step_time 1398.58, it/s 0.72
  1115. [0130 17:32:58 @multigpu.py:323] [p0574]  step: count(123), step_time 1531.42, mean_step_time 1439.55, it/s 0.69
  1116. [0130 17:32:59 @multigpu.py:323] [p0115]  step: count(104), step_time 1381.62, mean_step_time 1386.86, it/s 0.72
  1117. [0130 17:32:59 @multigpu.py:323] [p0576]  step: count(126), step_time 1381.94, mean_step_time 1397.42, it/s 0.72
  1118. [0130 17:33:00 @multigpu.py:323] [p0574]  step: count(124), step_time 1630.48, mean_step_time 1452.51, it/s 0.69
  1119. [0130 17:33:00 @multigpu.py:323] [p0115]  step: count(105), step_time 1369.54, mean_step_time 1385.63, it/s 0.72
  1120. [0130 17:33:00 @multigpu.py:323] [p0576]  step: count(127), step_time 1395.76, mean_step_time 1397.63, it/s 0.72
  1121. [0130 17:33:02 @multigpu.py:323] [p0115]  step: count(106), step_time 1406.69, mean_step_time 1386.17, it/s 0.72
  1122. [0130 17:33:02 @multigpu.py:323] [p0574]  step: count(125), step_time 1546.76, mean_step_time 1458.92, it/s 0.69
  1123. [0130 17:33:02 @multigpu.py:323] [p0576]  step: count(128), step_time 1399.48, mean_step_time 1395.2, it/s 0.72
  1124. [0130 17:33:03 @multigpu.py:323] [p0576]  step: count(129), step_time 1363.1, mean_step_time 1391.25, it/s 0.72
  1125. [0130 17:33:04 @multigpu.py:323] [p0576]  step: count(130), step_time 1338.32, mean_step_time 1390.72, it/s 0.72
  1126. [0130 17:33:04 @multigpu.py:323] [p0115]  step: count(107), step_time 2796.9, mean_step_time 1458.32, it/s 0.69
  1127. [0130 17:33:04 @multigpu.py:323] [p0574]  step: count(126), step_time 2779.41, mean_step_time 1529.82, it/s 0.65
  1128. sending to address tcp://p0112:61216
  1129. ##### Sending to neptune:  online_score :  0.0510955774784 , 1.3 #####
  1130. [u'online', 1.3]
  1131. receiving
  1132. [0130 17:33:06 @multigpu.py:323] [p0115]  step: count(108), step_time 1399.23, mean_step_time 1457.67, it/s 0.69
  1133. [0130 17:33:06 @multigpu.py:323] [p0576]  step: count(131), step_time 1472.37, mean_step_time 1395.44, it/s 0.72
  1134. [0130 17:33:06 @multigpu.py:323] [p0574]  step: count(127), step_time 1573.33, mean_step_time 1537.01, it/s 0.65
  1135. [0130 17:33:07 @multigpu.py:323] [p0115]  step: count(109), step_time 1380.55, mean_step_time 1458.17, it/s 0.69
  1136. [0130 17:33:07 @multigpu.py:323] [p0576]  step: count(132), step_time 1427.56, mean_step_time 1397.49, it/s 0.72
  1137. [0130 17:33:08 @multigpu.py:323] [p0574]  step: count(128), step_time 1619.61, mean_step_time 1548.65, it/s 0.65
  1138. [0130 17:33:09 @multigpu.py:323] [p0115]  step: count(110), step_time 1370.05, mean_step_time 1459.08, it/s 0.69
  1139. [0130 17:33:09 @multigpu.py:323] [p0576]  step: count(133), step_time 1399.82, mean_step_time 1396.12, it/s 0.72
  1140. sending to address tcp://p0112:61216
  1141. ##### Sending to neptune:  online_score :  0.0522849238581 , 1.6 #####
  1142. [u'online', 1.6]
  1143. ##### Sending to neptune:  active_workers :  0.0522849566407 , 3 #####
  1144. receiving
  1145. [0130 17:33:09 @multigpu.py:323] [p0574]  step: count(129), step_time 1624.04, mean_step_time 1559.29, it/s 0.64
  1146. [0130 17:33:10 @multigpu.py:323] [p0115]  step: count(111), step_time 1376.21, mean_step_time 1459.9, it/s 0.68
  1147. [0130 17:33:10 @multigpu.py:323] [p0576]  step: count(134), step_time 1361.15, mean_step_time 1396.61, it/s 0.72
  1148. [0130 17:33:11 @multigpu.py:323] [p0574]  step: count(130), step_time 1590.44, mean_step_time 1566.17, it/s 0.64
  1149. [0130 17:33:11 @multigpu.py:323] [p0115]  step: count(112), step_time 1365.93, mean_step_time 1455.38, it/s 0.69
  1150. [0130 17:33:11 @multigpu.py:323] [p0576]  step: count(135), step_time 1388.74, mean_step_time 1393.05, it/s 0.72
  1151. [0130 17:33:13 @multigpu.py:323] [p0574]  step: count(131), step_time 1731.61, mean_step_time 1585.06, it/s 0.63
  1152. [0130 17:33:13 @multigpu.py:323] [p0115]  step: count(113), step_time 1383.15, mean_step_time 1456.34, it/s 0.69
  1153. [0130 17:33:13 @multigpu.py:323] [p0576]  step: count(136), step_time 1343.67, mean_step_time 1389.17, it/s 0.72
  1154. [0130 17:33:14 @multigpu.py:323] [p0115]  step: count(114), step_time 1378.8, mean_step_time 1456.78, it/s 0.69
  1155. [0130 17:33:14 @multigpu.py:323] [p0574]  step: count(132), step_time 1601.35, mean_step_time 1594.7, it/s 0.63
  1156. [0130 17:33:14 @multigpu.py:323] [p0576]  step: count(137), step_time 1405.46, mean_step_time 1390.31, it/s 0.72
  1157. [0130 17:33:15 @multigpu.py:323] [p0115]  step: count(115), step_time 1399.48, mean_step_time 1457.81, it/s 0.69
  1158. [0130 17:33:16 @multigpu.py:323] [p0576]  step: count(138), step_time 1388.43, mean_step_time 1391.27, it/s 0.72
  1159. [0130 17:33:16 @multigpu.py:323] [p0574]  step: count(133), step_time 1626.81, mean_step_time 1604.41, it/s 0.62
  1160. [0130 17:33:17 @multigpu.py:323] [p0115]  step: count(116), step_time 1407.25, mean_step_time 1458.08, it/s 0.69
  1161. [0130 17:33:17 @multigpu.py:323] [p0576]  step: count(139), step_time 1346.21, mean_step_time 1389.11, it/s 0.72
  1162. [0130 17:33:17 @multigpu.py:323] [p0574]  step: count(134), step_time 1527.52, mean_step_time 1609.13, it/s 0.62
  1163. [0130 17:33:18 @multigpu.py:323] [p0115]  step: count(117), step_time 1409.14, mean_step_time 1459.43, it/s 0.69
  1164. [0130 17:33:18 @multigpu.py:323] [p0576]  step: count(140), step_time 1372.22, mean_step_time 1386.23, it/s 0.72
  1165. [0130 17:33:19 @multigpu.py:323] [p0574]  step: count(135), step_time 1426.88, mean_step_time 1608.95, it/s 0.62
  1166. [0130 17:33:20 @multigpu.py:323] [p0115]  step: count(118), step_time 1375.3, mean_step_time 1456.01, it/s 0.69
  1167. [0130 17:33:20 @multigpu.py:323] [p0576]  step: count(141), step_time 1366.67, mean_step_time 1384.91, it/s 0.72
  1168. [0130 17:33:20 @multigpu.py:323] [p0574]  step: count(136), step_time 1466.03, mean_step_time 1611.19, it/s 0.62
  1169. [0130 17:33:21 @multigpu.py:323] [p0576]  step: count(142), step_time 1389.33, mean_step_time 1384.76, it/s 0.72
  1170. [0130 17:33:21 @multigpu.py:323] [p0115]  step: count(119), step_time 1424.17, mean_step_time 1457.83, it/s 0.69
  1171. [0130 17:33:22 @multigpu.py:323] [p0574]  step: count(137), step_time 1382.97, mean_step_time 1610.27, it/s 0.62
  1172. [0130 17:33:22 @multigpu.py:323] [p0576]  step: count(143), step_time 1315.06, mean_step_time 1383.09, it/s 0.72
  1173. [0130 17:33:22 @multigpu.py:323] [p0115]  step: count(120), step_time 1378.73, mean_step_time 1457.7, it/s 0.69
  1174. [0130 17:33:23 @multigpu.py:323] [p0574]  step: count(138), step_time 1377.13, mean_step_time 1606.41, it/s 0.62
  1175. [0130 17:33:24 @multigpu.py:323] [p0576]  step: count(144), step_time 1410.72, mean_step_time 1385.46, it/s 0.72
  1176. [0130 17:33:24 @multigpu.py:323] [p0115]  step: count(121), step_time 1428.49, mean_step_time 1458.61, it/s 0.69
  1177. [0130 17:33:24 @multigpu.py:323] [p0574]  step: count(139), step_time 1379.38, mean_step_time 1599.02, it/s 0.63
  1178. sending to address tcp://p0112:61216
  1179. ##### Sending to neptune:  online_score :  0.0566255952252 , 1.6 #####
  1180. [u'online', 1.6]
  1181. receiving
  1182. sending to address tcp://p0112:61216
  1183. ##### Sending to neptune:  online_score :  0.0567920777533 , 1.3 #####
  1184. [u'online', 1.3]
  1185. receiving
  1186. [0130 17:33:25 @multigpu.py:323] [p0576]  step: count(145), step_time 1331.44, mean_step_time 1379.87, it/s 0.72
  1187. [0130 17:33:25 @multigpu.py:323] [p0115]  step: count(122), step_time 1338.18, mean_step_time 1457.65, it/s 0.69
  1188. [0130 17:33:26 @multigpu.py:323] [p0574]  step: count(140), step_time 1408.26, mean_step_time 1592.89, it/s 0.63
  1189. [0130 17:33:27 @multigpu.py:323] [p0576]  step: count(146), step_time 1414.04, mean_step_time 1381.48, it/s 0.72
  1190. [0130 17:33:27 @multigpu.py:323] [p0115]  step: count(123), step_time 1422.33, mean_step_time 1459.59, it/s 0.69
  1191. [0130 17:33:27 @multigpu.py:323] [p0574]  step: count(141), step_time 1368.26, mean_step_time 1588.68, it/s 0.63
  1192. [0130 17:33:28 @multigpu.py:323] [p0576]  step: count(147), step_time 1339.09, mean_step_time 1378.64, it/s 0.73
  1193. [0130 17:33:28 @multigpu.py:323] [p0115]  step: count(124), step_time 1376.01, mean_step_time 1459.31, it/s 0.69
  1194. [0130 17:33:29 @multigpu.py:323] [p0574]  step: count(142), step_time 1417.7, mean_step_time 1580.47, it/s 0.63
  1195. [0130 17:33:29 @multigpu.py:323] [p0576]  step: count(148), step_time 1386.21, mean_step_time 1377.98, it/s 0.73
  1196. [0130 17:33:29 @multigpu.py:323] [p0115]  step: count(125), step_time 1357.21, mean_step_time 1458.69, it/s 0.69
  1197. [0130 17:33:30 @multigpu.py:323] [p0574]  step: count(143), step_time 1351.42, mean_step_time 1571.47, it/s 0.64
  1198. [0130 17:33:31 @multigpu.py:323] [p0576]  step: count(149), step_time 1354.23, mean_step_time 1377.54, it/s 0.73
  1199. [0130 17:33:31 @multigpu.py:323] [p0115]  step: count(126), step_time 1383.09, mean_step_time 1457.51, it/s 0.69
  1200. [0130 17:33:31 @multigpu.py:323] [p0574]  step: count(144), step_time 1374.31, mean_step_time 1558.66, it/s 0.64
  1201. [0130 17:33:32 @multigpu.py:323] [p0576]  step: count(150), step_time 1377.74, mean_step_time 1379.51, it/s 0.72
  1202. [0130 17:33:33 @multigpu.py:323] [p0576]  step: count(151), step_time 1383.28, mean_step_time 1375.05, it/s 0.73
  1203. [0130 17:33:33 @multigpu.py:323] [p0574]  step: count(145), step_time 2126.63, mean_step_time 1587.65, it/s 0.63
  1204. [0130 17:33:33 @multigpu.py:323] [p0115]  step: count(127), step_time 2620.68, mean_step_time 1448.7, it/s 0.69
  1205. sending to address tcp://p0112:61216
  1206. ##### Sending to neptune:  online_score :  0.0592391738627 , 1.6 #####
  1207. [u'online', 1.6]
  1208. receiving
  1209. [0130 17:33:35 @multigpu.py:323] [p0574]  step: count(146), step_time 1375.08, mean_step_time 1517.44, it/s 0.66
  1210. [0130 17:33:35 @multigpu.py:323] [p0576]  step: count(152), step_time 1375.58, mean_step_time 1372.45, it/s 0.73
  1211. [0130 17:33:35 @multigpu.py:323] [p0115]  step: count(128), step_time 1440.45, mean_step_time 1450.76, it/s 0.69
  1212. [0130 17:33:36 @multigpu.py:323] [p0576]  step: count(153), step_time 1407.13, mean_step_time 1372.82, it/s 0.73
  1213. [0130 17:33:36 @multigpu.py:323] [p0115]  step: count(129), step_time 1366.56, mean_step_time 1450.06, it/s 0.69
  1214. [0130 17:33:36 @multigpu.py:323] [p0574]  step: count(147), step_time 1436.0, mean_step_time 1510.57, it/s 0.66
  1215. [0130 17:33:37 @multigpu.py:323] [p0576]  step: count(154), step_time 1322.97, mean_step_time 1370.91, it/s 0.73
  1216. [0130 17:33:38 @multigpu.py:323] [p0574]  step: count(148), step_time 1352.32, mean_step_time 1497.21, it/s 0.67
  1217. [0130 17:33:38 @multigpu.py:323] [p0115]  step: count(130), step_time 1406.66, mean_step_time 1451.89, it/s 0.69
  1218. [0130 17:33:39 @multigpu.py:323] [p0576]  step: count(155), step_time 1419.88, mean_step_time 1372.47, it/s 0.73
  1219. [0130 17:33:39 @multigpu.py:323] [p0574]  step: count(149), step_time 1408.69, mean_step_time 1486.44, it/s 0.67
  1220. [0130 17:33:39 @multigpu.py:323] [p0115]  step: count(131), step_time 1452.33, mean_step_time 1455.7, it/s 0.69
  1221. [0130 17:33:40 @multigpu.py:323] [p0576]  step: count(156), step_time 1347.77, mean_step_time 1372.67, it/s 0.73
  1222. [0130 17:33:40 @multigpu.py:323] [p0574]  step: count(150), step_time 1384.36, mean_step_time 1476.13, it/s 0.68
  1223. [0130 17:33:40 @multigpu.py:323] [p0115]  step: count(132), step_time 1370.15, mean_step_time 1455.91, it/s 0.69
  1224. [0130 17:33:42 @multigpu.py:323] [p0576]  step: count(157), step_time 1357.76, mean_step_time 1370.29, it/s 0.73
  1225. [0130 17:33:42 @multigpu.py:323] [p0574]  step: count(151), step_time 1339.38, mean_step_time 1456.52, it/s 0.69
  1226. [0130 17:33:42 @multigpu.py:323] [p0115]  step: count(133), step_time 1349.21, mean_step_time 1454.21, it/s 0.69
  1227. [0130 17:33:43 @multigpu.py:323] [p0576]  step: count(158), step_time 1393.96, mean_step_time 1370.56, it/s 0.73
  1228. [0130 17:33:43 @multigpu.py:323] [p0574]  step: count(152), step_time 1389.87, mean_step_time 1445.95, it/s 0.69
  1229. [0130 17:33:43 @multigpu.py:323] [p0115]  step: count(134), step_time 1368.66, mean_step_time 1453.7, it/s 0.69
  1230. sending to address tcp://p0112:61216
  1231. ##### Sending to neptune:  online_score :  0.0620354730553 , 1.4 #####
  1232. [u'online', 1.4]
  1233. receiving
  1234. [0130 17:33:44 @multigpu.py:323] [p0576]  step: count(159), step_time 1365.54, mean_step_time 1371.53, it/s 0.73
  1235. [0130 17:33:44 @multigpu.py:323] [p0574]  step: count(153), step_time 1363.79, mean_step_time 1432.8, it/s 0.7
  1236. [0130 17:33:44 @multigpu.py:323] [p0115]  step: count(135), step_time 1364.54, mean_step_time 1451.96, it/s 0.69
  1237. [0130 17:33:46 @multigpu.py:323] [p0576]  step: count(160), step_time 1351.72, mean_step_time 1370.5, it/s 0.73
  1238. [0130 17:33:46 @multigpu.py:323] [p0574]  step: count(154), step_time 1415.48, mean_step_time 1427.2, it/s 0.7
  1239. [0130 17:33:46 @multigpu.py:323] [p0115]  step: count(136), step_time 1433.35, mean_step_time 1453.26, it/s 0.69
  1240. [0130 17:33:47 @multigpu.py:323] [p0576]  step: count(161), step_time 1397.89, mean_step_time 1372.07, it/s 0.73
  1241. [0130 17:33:47 @multigpu.py:323] [p0574]  step: count(155), step_time 1358.27, mean_step_time 1423.77, it/s 0.7
  1242. [0130 17:33:47 @multigpu.py:323] [p0115]  step: count(137), step_time 1386.5, mean_step_time 1452.13, it/s 0.69
  1243. [0130 17:33:48 @multigpu.py:323] [p0576]  step: count(162), step_time 1338.21, mean_step_time 1369.51, it/s 0.73
  1244. [0130 17:33:49 @multigpu.py:323] [p0574]  step: count(156), step_time 1386.11, mean_step_time 1419.77, it/s 0.7
  1245. [0130 17:33:49 @multigpu.py:323] [p0115]  step: count(138), step_time 1378.19, mean_step_time 1452.27, it/s 0.69
  1246. [0130 17:33:50 @multigpu.py:323] [p0576]  step: count(163), step_time 1400.41, mean_step_time 1373.78, it/s 0.73
  1247. [0130 17:33:50 @multigpu.py:323] [p0574]  step: count(157), step_time 1369.1, mean_step_time 1419.08, it/s 0.7
  1248. [0130 17:33:50 @multigpu.py:323] [p0115]  step: count(139), step_time 1378.65, mean_step_time 1450.0, it/s 0.69
  1249. [0130 17:33:51 @multigpu.py:323] [p0576]  step: count(164), step_time 1331.14, mean_step_time 1369.8, it/s 0.73
  1250. [0130 17:33:51 @multigpu.py:323] [p0574]  step: count(158), step_time 1411.4, mean_step_time 1420.79, it/s 0.7
  1251. [0130 17:33:51 @multigpu.py:323] [p0115]  step: count(140), step_time 1361.03, mean_step_time 1449.11, it/s 0.69
  1252. [0130 17:33:53 @multigpu.py:323] [p0576]  step: count(165), step_time 1387.46, mean_step_time 1372.6, it/s 0.73
  1253. [0130 17:33:53 @multigpu.py:323] [p0115]  step: count(141), step_time 1417.65, mean_step_time 1448.57, it/s 0.69
  1254. [0130 17:33:53 @multigpu.py:323] [p0574]  step: count(159), step_time 1508.85, mean_step_time 1427.26, it/s 0.7
  1255. [0130 17:33:54 @multigpu.py:323] [p0576]  step: count(166), step_time 1382.32, mean_step_time 1371.01, it/s 0.73
  1256. [0130 17:33:54 @multigpu.py:323] [p0115]  step: count(142), step_time 1374.83, mean_step_time 1450.4, it/s 0.69
  1257. [0130 17:33:54 @multigpu.py:323] [p0574]  step: count(160), step_time 1371.93, mean_step_time 1425.45, it/s 0.7
  1258. [0130 17:33:55 @multigpu.py:323] [p0576]  step: count(167), step_time 1386.89, mean_step_time 1373.4, it/s 0.73
  1259. [0130 17:33:56 @multigpu.py:323] [p0115]  step: count(143), step_time 1362.51, mean_step_time 1447.41, it/s 0.69
  1260. [0130 17:33:56 @multigpu.py:323] [p0574]  step: count(161), step_time 1391.61, mean_step_time 1426.61, it/s 0.7
  1261. [0130 17:33:57 @multigpu.py:323] [p0576]  step: count(168), step_time 1337.04, mean_step_time 1370.95, it/s 0.73
  1262. [0130 17:33:57 @multigpu.py:323] [p0574]  step: count(162), step_time 1362.97, mean_step_time 1423.88, it/s 0.7
  1263. [0130 17:33:57 @multigpu.py:323] [p0115]  step: count(144), step_time 1451.4, mean_step_time 1451.18, it/s 0.69
  1264. sending to address tcp://p0112:61216
  1265. ##### Sending to neptune:  online_score :  0.0657542652554 , 0.9 #####
  1266. [u'online', 0.9]
  1267. receiving
  1268. [0130 17:33:58 @multigpu.py:323] [p0576]  step: count(169), step_time 1383.69, mean_step_time 1372.42, it/s 0.73
  1269. [0130 17:33:58 @multigpu.py:323] [p0574]  step: count(163), step_time 1399.8, mean_step_time 1426.3, it/s 0.7
  1270. [0130 17:33:58 @multigpu.py:323] [p0115]  step: count(145), step_time 1387.85, mean_step_time 1452.71, it/s 0.69
  1271. sending to address tcp://p0112:61216
  1272. ##### Sending to neptune:  online_score :  0.0660777313842 , 1.3 #####
  1273. [u'online', 1.3]
  1274. receiving
  1275. [0130 17:33:59 @multigpu.py:323] [p0576]  step: count(170), step_time 1392.35, mean_step_time 1373.15, it/s 0.73
  1276. [0130 17:34:00 @multigpu.py:323] [p0115]  step: count(146), step_time 1359.71, mean_step_time 1451.55, it/s 0.69
  1277. [0130 17:34:00 @multigpu.py:323] [p0574]  step: count(164), step_time 1395.34, mean_step_time 1427.35, it/s 0.7
  1278. [0130 17:34:01 @multigpu.py:323] [p0115]  step: count(147), step_time 1373.86, mean_step_time 1389.2, it/s 0.72
  1279. [0130 17:34:01 @multigpu.py:323] [p0574]  step: count(165), step_time 1364.25, mean_step_time 1389.23, it/s 0.72
  1280. [0130 17:34:01 @multigpu.py:323] [p0576]  step: count(171), step_time 1705.46, mean_step_time 1389.26, it/s 0.72
  1281. [0130 17:34:03 @multigpu.py:323] [p0115]  step: count(148), step_time 1351.29, mean_step_time 1384.75, it/s 0.72
  1282. [0130 17:34:03 @multigpu.py:323] [p0576]  step: count(172), step_time 1396.38, mean_step_time 1390.3, it/s 0.72
  1283. [0130 17:34:03 @multigpu.py:323] [p0574]  step: count(166), step_time 1407.4, mean_step_time 1390.85, it/s 0.72
  1284. [0130 17:34:04 @multigpu.py:323] [p0115]  step: count(149), step_time 1391.96, mean_step_time 1386.02, it/s 0.72
  1285. [0130 17:34:04 @multigpu.py:323] [p0574]  step: count(167), step_time 1387.59, mean_step_time 1388.43, it/s 0.72
  1286. [0130 17:34:04 @multigpu.py:323] [p0576]  step: count(173), step_time 1418.44, mean_step_time 1390.86, it/s 0.72
  1287. [0130 17:34:05 @multigpu.py:323] [p0115]  step: count(150), step_time 1357.6, mean_step_time 1383.56, it/s 0.72
  1288. [0130 17:34:05 @multigpu.py:323] [p0574]  step: count(168), step_time 1366.03, mean_step_time 1389.11, it/s 0.72
  1289. [0130 17:34:05 @multigpu.py:323] [p0576]  step: count(174), step_time 1388.43, mean_step_time 1394.14, it/s 0.72
  1290. [0130 17:34:07 @multigpu.py:323] [p0115]  step: count(151), step_time 1400.82, mean_step_time 1380.99, it/s 0.72
  1291. [0130 17:34:07 @multigpu.py:323] [p0574]  step: count(169), step_time 1367.17, mean_step_time 1387.04, it/s 0.72
  1292. [0130 17:34:07 @multigpu.py:323] [p0576]  step: count(175), step_time 1413.54, mean_step_time 1393.82, it/s 0.72
  1293. [0130 17:34:08 @multigpu.py:323] [p0115]  step: count(152), step_time 1361.39, mean_step_time 1380.55, it/s 0.72
  1294. [0130 17:34:08 @multigpu.py:323] [p0574]  step: count(170), step_time 1396.26, mean_step_time 1387.63, it/s 0.72
  1295. [0130 17:34:08 @multigpu.py:323] [p0576]  step: count(176), step_time 1380.22, mean_step_time 1395.44, it/s 0.72
  1296. [0130 17:34:09 @multigpu.py:323] [p0115]  step: count(153), step_time 1385.38, mean_step_time 1382.36, it/s 0.72
  1297. [0130 17:34:09 @multigpu.py:323] [p0574]  step: count(171), step_time 1366.67, mean_step_time 1388.99, it/s 0.72
  1298. [0130 17:34:10 @multigpu.py:323] [p0576]  step: count(177), step_time 1360.44, mean_step_time 1395.58, it/s 0.72
  1299. sending to address tcp://p0112:61216
  1300. ##### Sending to neptune:  online_score :  0.0692923605442 , 0.7 #####
  1301. [u'online', 0.7]
  1302. ##### Sending to neptune:  active_workers :  0.0692924124665 , 3 #####
  1303. receiving
  1304. [0130 17:34:11 @multigpu.py:323] [p0115]  step: count(154), step_time 1383.39, mean_step_time 1383.09, it/s 0.72
  1305. [0130 17:34:11 @multigpu.py:323] [p0574]  step: count(172), step_time 1391.29, mean_step_time 1389.07, it/s 0.72
  1306. [0130 17:34:11 @multigpu.py:323] [p0576]  step: count(178), step_time 1413.78, mean_step_time 1396.57, it/s 0.72
  1307. [0130 17:34:12 @multigpu.py:323] [p0115]  step: count(155), step_time 1360.79, mean_step_time 1382.91, it/s 0.72
  1308. [0130 17:34:12 @multigpu.py:323] [p0574]  step: count(173), step_time 1400.33, mean_step_time 1390.89, it/s 0.72
  1309. [0130 17:34:12 @multigpu.py:323] [p0576]  step: count(179), step_time 1355.16, mean_step_time 1396.05, it/s 0.72
  1310. [0130 17:34:14 @multigpu.py:323] [p0115]  step: count(156), step_time 1419.04, mean_step_time 1382.19, it/s 0.72
  1311. [0130 17:34:14 @multigpu.py:323] [p0574]  step: count(174), step_time 1371.01, mean_step_time 1388.67, it/s 0.72
  1312. [0130 17:34:14 @multigpu.py:323] [p0576]  step: count(180), step_time 1350.09, mean_step_time 1395.97, it/s 0.72
  1313. [0130 17:34:15 @multigpu.py:323] [p0115]  step: count(157), step_time 1368.62, mean_step_time 1381.3, it/s 0.72
  1314. [0130 17:34:15 @multigpu.py:323] [p0574]  step: count(175), step_time 1396.0, mean_step_time 1390.56, it/s 0.72
  1315. [0130 17:34:15 @multigpu.py:323] [p0576]  step: count(181), step_time 1375.06, mean_step_time 1394.83, it/s 0.72
  1316. [0130 17:34:16 @multigpu.py:323] [p0115]  step: count(158), step_time 1412.57, mean_step_time 1383.02, it/s 0.72
  1317. [0130 17:34:16 @multigpu.py:323] [p0574]  step: count(176), step_time 1350.15, mean_step_time 1388.76, it/s 0.72
  1318. [0130 17:34:16 @multigpu.py:323] [p0576]  step: count(182), step_time 1365.4, mean_step_time 1396.19, it/s 0.72
  1319. [0130 17:34:18 @multigpu.py:323] [p0115]  step: count(159), step_time 1386.58, mean_step_time 1383.41, it/s 0.72
  1320. [0130 17:34:18 @multigpu.py:323] [p0576]  step: count(183), step_time 1362.43, mean_step_time 1394.29, it/s 0.72
  1321. [0130 17:34:18 @multigpu.py:323] [p0574]  step: count(177), step_time 1407.9, mean_step_time 1390.7, it/s 0.72
  1322. [0130 17:34:19 @multigpu.py:323] [p0576]  step: count(184), step_time 1366.59, mean_step_time 1396.06, it/s 0.72
  1323. [0130 17:34:19 @multigpu.py:323] [p0115]  step: count(160), step_time 1374.1, mean_step_time 1384.07, it/s 0.72
  1324. [0130 17:34:19 @multigpu.py:323] [p0574]  step: count(178), step_time 1409.13, mean_step_time 1390.58, it/s 0.72
  1325. [0130 17:34:20 @multigpu.py:323] [p0576]  step: count(185), step_time 1375.44, mean_step_time 1395.46, it/s 0.72
  1326. [0130 17:34:20 @multigpu.py:323] [p0574]  step: count(179), step_time 1321.83, mean_step_time 1381.23, it/s 0.72
  1327. [0130 17:34:21 @multigpu.py:323] [p0115]  step: count(161), step_time 1406.45, mean_step_time 1383.51, it/s 0.72
  1328. [0130 17:34:22 @multigpu.py:323] [p0576]  step: count(186), step_time 1394.41, mean_step_time 1396.06, it/s 0.72
  1329. [0130 17:34:22 @multigpu.py:323] [p0574]  step: count(180), step_time 1405.91, mean_step_time 1382.93, it/s 0.72
  1330. [0130 17:34:22 @multigpu.py:323] [p0115]  step: count(162), step_time 1402.49, mean_step_time 1384.89, it/s 0.72
  1331. [0130 17:34:23 @multigpu.py:323] [p0576]  step: count(187), step_time 1335.38, mean_step_time 1393.49, it/s 0.72
  1332. [0130 17:34:23 @multigpu.py:323] [p0115]  step: count(163), step_time 1373.02, mean_step_time 1385.41, it/s 0.72
  1333. [0130 17:34:23 @multigpu.py:323] [p0574]  step: count(181), step_time 1410.48, mean_step_time 1383.88, it/s 0.72
  1334. [0130 17:34:25 @multigpu.py:323] [p0576]  step: count(188), step_time 1389.98, mean_step_time 1396.13, it/s 0.72
  1335. [0130 17:34:25 @multigpu.py:323] [p0574]  step: count(182), step_time 1353.91, mean_step_time 1383.42, it/s 0.72
  1336. [0130 17:34:25 @multigpu.py:323] [p0115]  step: count(164), step_time 1392.64, mean_step_time 1382.48, it/s 0.72
  1337. sending to address tcp://p0112:61216
  1338. ##### Sending to neptune:  online_score :  0.0735821829902 , 1.7 #####
  1339. [u'online', 1.7]
  1340. receiving
  1341. [0130 17:34:26 @multigpu.py:323] [p0576]  step: count(189), step_time 1328.79, mean_step_time 1393.39, it/s 0.72
  1342. [0130 17:34:26 @multigpu.py:323] [p0115]  step: count(165), step_time 1361.66, mean_step_time 1381.17, it/s 0.72
  1343. [0130 17:34:26 @multigpu.py:323] [p0574]  step: count(183), step_time 1382.06, mean_step_time 1382.54, it/s 0.72
  1344. [0130 17:34:27 @multigpu.py:323] [p0576]  step: count(190), step_time 1381.09, mean_step_time 1392.83, it/s 0.72
  1345. [0130 17:34:27 @multigpu.py:323] [p0574]  step: count(184), step_time 1377.0, mean_step_time 1381.62, it/s 0.72
  1346. [0130 17:34:27 @multigpu.py:323] [p0115]  step: count(166), step_time 1383.95, mean_step_time 1382.38, it/s 0.72
  1347. [0130 17:34:29 @multigpu.py:323] [p0115]  step: count(167), step_time 1391.89, mean_step_time 1383.28, it/s 0.72
  1348. [0130 17:34:29 @multigpu.py:323] [p0574]  step: count(185), step_time 1398.33, mean_step_time 1383.32, it/s 0.72
  1349. [0130 17:34:29 @multigpu.py:323] [p0576]  step: count(191), step_time 1513.07, mean_step_time 1383.21, it/s 0.72
  1350. [0130 17:34:30 @multigpu.py:323] [p0115]  step: count(168), step_time 1358.74, mean_step_time 1383.65, it/s 0.72
  1351. [0130 17:34:30 @multigpu.py:323] [p0574]  step: count(186), step_time 1370.96, mean_step_time 1381.5, it/s 0.72
  1352. [0130 17:34:30 @multigpu.py:323] [p0576]  step: count(192), step_time 1385.18, mean_step_time 1382.65, it/s 0.72
  1353. [0130 17:34:32 @multigpu.py:323] [p0576]  step: count(193), step_time 1358.67, mean_step_time 1379.66, it/s 0.72
  1354. sending to address tcp://p0112:61216
  1355. ##### Sending to neptune:  online_score :  0.0752706305186 , 0.8 #####
  1356. [u'online', 0.8]
  1357. receiving
  1358. [0130 17:34:32 @multigpu.py:323] [p0574]  step: count(187), step_time 1392.0, mean_step_time 1381.72, it/s 0.72
  1359. [0130 17:34:32 @multigpu.py:323] [p0115]  step: count(169), step_time 1417.02, mean_step_time 1384.91, it/s 0.72
  1360. [0130 17:34:33 @multigpu.py:323] [p0115]  step: count(170), step_time 1332.3, mean_step_time 1383.64, it/s 0.72
  1361. [0130 17:34:33 @multigpu.py:323] [p0574]  step: count(188), step_time 1353.15, mean_step_time 1381.08, it/s 0.72
  1362. [0130 17:34:33 @multigpu.py:323] [p0576]  step: count(194), step_time 1381.28, mean_step_time 1379.3, it/s 0.73
  1363. [0130 17:34:34 @multigpu.py:323] [p0576]  step: count(195), step_time 1359.54, mean_step_time 1376.6, it/s 0.73
  1364. [0130 17:34:34 @multigpu.py:323] [p0574]  step: count(189), step_time 1392.51, mean_step_time 1382.34, it/s 0.72
  1365. [0130 17:34:34 @multigpu.py:323] [p0115]  step: count(171), step_time 1418.38, mean_step_time 1384.52, it/s 0.72
  1366. [0130 17:34:36 @multigpu.py:323] [p0576]  step: count(196), step_time 1368.99, mean_step_time 1376.04, it/s 0.73
  1367. [0130 17:34:36 @multigpu.py:323] [p0574]  step: count(190), step_time 1355.72, mean_step_time 1380.32, it/s 0.72
  1368. [0130 17:34:36 @multigpu.py:323] [p0115]  step: count(172), step_time 1368.87, mean_step_time 1384.89, it/s 0.72
  1369. [0130 17:34:37 @multigpu.py:323] [p0576]  step: count(197), step_time 1351.55, mean_step_time 1375.59, it/s 0.73
  1370. [0130 17:34:37 @multigpu.py:323] [p0574]  step: count(191), step_time 1382.27, mean_step_time 1381.1, it/s 0.72
  1371. [0130 17:34:37 @multigpu.py:323] [p0115]  step: count(173), step_time 1384.74, mean_step_time 1384.86, it/s 0.72
  1372. sending to address tcp://p0112:61216
  1373. ##### Sending to neptune:  online_score :  0.0769521447023 , 1.2 #####
  1374. [u'online', 1.2]
  1375. receiving
  1376. [0130 17:34:38 @multigpu.py:323] [p0576]  step: count(198), step_time 1382.29, mean_step_time 1374.02, it/s 0.73
  1377. [0130 17:34:38 @multigpu.py:323] [p0574]  step: count(192), step_time 1367.78, mean_step_time 1379.92, it/s 0.72
  1378. [0130 17:34:38 @multigpu.py:323] [p0115]  step: count(174), step_time 1381.63, mean_step_time 1384.77, it/s 0.72
  1379. [0130 17:34:40 @multigpu.py:323] [p0576]  step: count(199), step_time 1398.43, mean_step_time 1376.18, it/s 0.73
  1380. [0130 17:34:40 @multigpu.py:323] [p0574]  step: count(193), step_time 1392.39, mean_step_time 1379.52, it/s 0.72
  1381. [0130 17:34:40 @multigpu.py:323] [p0115]  step: count(175), step_time 1386.47, mean_step_time 1386.06, it/s 0.72
  1382. [0130 17:34:41 @multigpu.py:323] [p0576]  step: count(200), step_time 1342.74, mean_step_time 1375.81, it/s 0.73
  1383. sending debugging info...
  1384. sending to address tcp://p0112:61216
  1385. sending to address tcp://p0112:61216
  1386. ##### Sending to neptune:  mean_delay :  0.0779323197073 , 0.0 #####
  1387. ##### Sending to neptune:  max_delay :  0.0779323197073 , -0.0 #####
  1388. ##### Sending to neptune:  min_delay :  0.0779323197073 , -0.0 #####
  1389. [u'delays', [0.0, -0.0, -0.0]]
  1390. receiving
  1391. ##### Sending to neptune:  cost :  0.0779328760836 , -0.00780312530696 #####
  1392. sending to address tcp://p0112:61216
  1393. ##### Sending to neptune:  policy_loss :  0.0779328760836 , 0.118199490011 #####
  1394. ##### Sending to neptune:  xentropy_loss :  0.0779328760836 , -2.29004001617 #####
  1395. ##### Sending to neptune:  value_loss :  0.0779328760836 , 1.17304027081 #####
  1396. ##### Sending to neptune:  advantage :  0.0779328760836 , -0.000583324115723 #####
  1397. ##### Sending to neptune:  pred_reward :  0.0779328760836 , 0.359293758869 #####
  1398. ##### Sending to neptune:  max_logit :  0.0779328760836 , 0.189353853464 #####
  1399. [u'loss', -0.007803125306963921, 0.11819949001073837, -2.2900400161743164, 1.1730402708053589, -0.0005833241157233715, 0.35929375886917114, 0.1893538534641266]
  1400. receiving
  1401. ##### Sending to neptune:  active_relus :  0.0779333224561 , 8862300.92 #####
  1402. ##### Sending to neptune:  dp_per_s :  0.0779333224561 , 92.2689357559 #####
  1403. [u'other', 8862300.92, 92.26893575586253]
  1404. receiving
  1405. [0130 17:34:41 @multigpu.py:323] [p0574]  step: count(194), step_time 1399.89, mean_step_time 1380.97, it/s 0.72
  1406. [0130 17:34:41 @multigpu.py:323] [p0115]  step: count(176), step_time 1364.66, mean_step_time 1383.34, it/s 0.72
  1407. [0130 17:34:43 @multigpu.py:323] [p0576]  step: count(201), step_time 1375.37, mean_step_time 1375.83, it/s 0.73
  1408. [0130 17:34:43 @multigpu.py:323] [p0115]  step: count(177), step_time 1421.45, mean_step_time 1385.98, it/s 0.72
  1409. [0130 17:34:43 @multigpu.py:323] [p0574]  step: count(195), step_time 1460.79, mean_step_time 1384.21, it/s 0.72
  1410. [0130 17:34:44 @multigpu.py:323] [p0576]  step: count(202), step_time 1389.25, mean_step_time 1377.02, it/s 0.73
  1411. [0130 17:34:44 @multigpu.py:323] [p0115]  step: count(178), step_time 1326.35, mean_step_time 1381.67, it/s 0.72
  1412. [0130 17:34:44 @multigpu.py:323] [p0574]  step: count(196), step_time 1371.53, mean_step_time 1385.28, it/s 0.72
  1413. [0130 17:34:45 @multigpu.py:323] [p0576]  step: count(203), step_time 1359.61, mean_step_time 1376.88, it/s 0.73
  1414. [0130 17:34:45 @multigpu.py:323] [p0115]  step: count(179), step_time 1378.42, mean_step_time 1381.26, it/s 0.72
  1415. [0130 17:34:46 @multigpu.py:323] [p0574]  step: count(197), step_time 1484.98, mean_step_time 1389.13, it/s 0.72
  1416. [0130 17:34:47 @multigpu.py:323] [p0576]  step: count(204), step_time 1380.84, mean_step_time 1377.59, it/s 0.73
  1417. [0130 17:34:47 @multigpu.py:323] [p0115]  step: count(180), step_time 1379.61, mean_step_time 1381.54, it/s 0.72
  1418. [0130 17:34:47 @multigpu.py:323] [p0574]  step: count(198), step_time 1512.61, mean_step_time 1394.3, it/s 0.72
  1419. [0130 17:34:48 @multigpu.py:323] [p0576]  step: count(205), step_time 1359.08, mean_step_time 1376.78, it/s 0.73
  1420. [0130 17:34:48 @multigpu.py:323] [p0115]  step: count(181), step_time 1355.4, mean_step_time 1378.99, it/s 0.73
  1421. [0130 17:34:48 @multigpu.py:323] [p0574]  step: count(199), step_time 1361.12, mean_step_time 1396.27, it/s 0.72
  1422. [0130 17:34:49 @multigpu.py:323] [p0576]  step: count(206), step_time 1383.91, mean_step_time 1376.25, it/s 0.73
  1423. [0130 17:34:50 @multigpu.py:323] [p0115]  step: count(182), step_time 1399.66, mean_step_time 1378.84, it/s 0.73
  1424. [0130 17:34:50 @multigpu.py:323] [p0574]  step: count(200), step_time 1434.78, mean_step_time 1397.71, it/s 0.72
  1425. sending debugging info...
  1426. sending to address tcp://p0112:61216
  1427. ##### Sending to neptune:  mean_delay :  0.0803513783216 , 0.0 #####
  1428. ##### Sending to neptune:  max_delay :  0.0803513783216 , -0.0 #####
  1429. ##### Sending to neptune:  min_delay :  0.0803513783216 , -0.0 #####
  1430. [u'delays', [0.0, -0.0, -0.0]]
  1431. sending to address tcp://p0112:61216
  1432. receiving
  1433. ##### Sending to neptune:  cost :  0.080351874431 , -0.00596463959664 #####
  1434. ##### Sending to neptune:  policy_loss :  0.080351874431 , 0.299598842859 #####
  1435. ##### Sending to neptune:  xentropy_loss :  0.080351874431 , -2.28987956047 #####
  1436. ##### Sending to neptune:  value_loss :  0.080351874431 , 1.22680675983 #####
  1437. sending to address tcp://p0112:61216
  1438. ##### Sending to neptune:  advantage :  0.080351874431 , -0.00136904569808 #####
  1439. ##### Sending to neptune:  pred_reward :  0.080351874431 , 0.362240970135 #####
  1440. ##### Sending to neptune:  max_logit :  0.080351874431 , 0.190133050084 #####
  1441. [u'loss', -0.005964639596641064, 0.2995988428592682, -2.289879560470581, 1.2268067598342896, -0.0013690456980839372, 0.3622409701347351, 0.19013305008411407]
  1442. receiving
  1443. ##### Sending to neptune:  active_relus :  0.0803523513344 , 8882302.82 #####
  1444. ##### Sending to neptune:  dp_per_s :  0.0803523513344 , 88.9797109993 #####
  1445. [u'other', 8882302.82, 88.97971099925284]
  1446. receiving
  1447. [0130 17:34:51 @multigpu.py:323] [p0576]  step: count(207), step_time 1366.96, mean_step_time 1377.83, it/s 0.73
  1448. [0130 17:34:51 @multigpu.py:323] [p0115]  step: count(183), step_time 1397.31, mean_step_time 1380.06, it/s 0.72
  1449. [0130 17:34:51 @multigpu.py:323] [p0574]  step: count(201), step_time 1405.29, mean_step_time 1397.45, it/s 0.72
  1450. sending to address tcp://p0112:61216
  1451. ##### Sending to neptune:  online_score :  0.0809522196982 , 1.4 #####
  1452. [u'online', 1.4]
  1453. receiving
  1454. [0130 17:34:52 @multigpu.py:323] [p0576]  step: count(208), step_time 1430.94, mean_step_time 1379.88, it/s 0.72
  1455. [0130 17:34:52 @multigpu.py:323] [p0115]  step: count(184), step_time 1378.93, mean_step_time 1379.37, it/s 0.72
  1456. [0130 17:34:53 @multigpu.py:323] [p0574]  step: count(202), step_time 1402.51, mean_step_time 1399.88, it/s 0.71
  1457. [0130 17:34:54 @multigpu.py:323] [p0576]  step: count(209), step_time 1332.2, mean_step_time 1380.05, it/s 0.72
  1458. [0130 17:34:54 @multigpu.py:323] [p0115]  step: count(185), step_time 1397.63, mean_step_time 1381.17, it/s 0.72
  1459. [0130 17:34:54 @multigpu.py:323] [p0574]  step: count(203), step_time 1388.81, mean_step_time 1400.22, it/s 0.71
  1460. [0130 17:34:55 @multigpu.py:323] [p0576]  step: count(210), step_time 1379.6, mean_step_time 1379.97, it/s 0.72
  1461. [0130 17:34:55 @multigpu.py:323] [p0115]  step: count(186), step_time 1369.04, mean_step_time 1380.43, it/s 0.72
  1462. [0130 17:34:55 @multigpu.py:323] [p0574]  step: count(204), step_time 1403.95, mean_step_time 1401.57, it/s 0.71
  1463. [0130 17:34:57 @multigpu.py:323] [p0574]  step: count(205), step_time 1470.88, mean_step_time 1405.2, it/s 0.71
  1464. [0130 17:34:57 @multigpu.py:323] [p0115]  step: count(187), step_time 1890.06, mean_step_time 1405.33, it/s 0.71
  1465. [0130 17:34:57 @multigpu.py:323] [p0576]  step: count(211), step_time 2018.29, mean_step_time 1405.24, it/s 0.71
  1466. sending to address tcp://p0112:61216
  1467. ##### Sending to neptune:  online_score :  0.0823914271593 , 1.3 #####
  1468. [u'online', 1.3]
  1469. receiving
  1470. [0130 17:34:58 @multigpu.py:323] [p0576]  step: count(212), step_time 1364.15, mean_step_time 1404.18, it/s 0.71
  1471. [0130 17:34:58 @multigpu.py:323] [p0115]  step: count(188), step_time 1385.2, mean_step_time 1406.66, it/s 0.71
  1472. [0130 17:34:58 @multigpu.py:323] [p0574]  step: count(206), step_time 1431.93, mean_step_time 1408.24, it/s 0.71
  1473. sending to address tcp://p0112:61216
  1474. ##### Sending to neptune:  online_score :  0.0828070782953 , 1.7 #####
  1475. [u'online', 1.7]
  1476. receiving
  1477. [0130 17:35:00 @multigpu.py:323] [p0576]  step: count(213), step_time 1372.01, mean_step_time 1404.85, it/s 0.71
  1478. [0130 17:35:00 @multigpu.py:323] [p0115]  step: count(189), step_time 1385.77, mean_step_time 1405.09, it/s 0.71
  1479. [0130 17:35:00 @multigpu.py:323] [p0574]  step: count(207), step_time 1425.89, mean_step_time 1409.94, it/s 0.71
  1480. [0130 17:35:01 @multigpu.py:323] [p0115]  step: count(190), step_time 1344.78, mean_step_time 1405.72, it/s 0.71
  1481. [0130 17:35:01 @multigpu.py:323] [p0576]  step: count(214), step_time 1387.69, mean_step_time 1405.17, it/s 0.71
  1482. [0130 17:35:01 @multigpu.py:323] [p0574]  step: count(208), step_time 1364.84, mean_step_time 1410.52, it/s 0.71
  1483. [0130 17:35:02 @multigpu.py:323] [p0576]  step: count(215), step_time 1356.17, mean_step_time 1405.0, it/s 0.71
  1484. [0130 17:35:02 @multigpu.py:323] [p0115]  step: count(191), step_time 1402.47, mean_step_time 1404.92, it/s 0.71
  1485. [0130 17:35:03 @multigpu.py:323] [p0574]  step: count(209), step_time 1413.0, mean_step_time 1411.55, it/s 0.71
  1486. [0130 17:35:04 @multigpu.py:323] [p0576]  step: count(216), step_time 1396.33, mean_step_time 1406.37, it/s 0.71
  1487. [0130 17:35:04 @multigpu.py:323] [p0115]  step: count(192), step_time 1363.41, mean_step_time 1404.65, it/s 0.71
  1488. [0130 17:35:04 @multigpu.py:323] [p0574]  step: count(210), step_time 1400.8, mean_step_time 1413.8, it/s 0.71
  1489. [0130 17:35:05 @multigpu.py:323] [p0576]  step: count(217), step_time 1388.98, mean_step_time 1408.24, it/s 0.71
  1490. [0130 17:35:05 @multigpu.py:323] [p0115]  step: count(193), step_time 1392.09, mean_step_time 1405.02, it/s 0.71
  1491. [0130 17:35:05 @multigpu.py:323] [p0574]  step: count(211), step_time 1408.62, mean_step_time 1415.12, it/s 0.71
  1492. [0130 17:35:07 @multigpu.py:323] [p0576]  step: count(218), step_time 1369.08, mean_step_time 1407.58, it/s 0.71
  1493. [0130 17:35:07 @multigpu.py:323] [p0115]  step: count(194), step_time 1404.0, mean_step_time 1406.14, it/s 0.71
  1494. [0130 17:35:07 @multigpu.py:323] [p0574]  step: count(212), step_time 1398.96, mean_step_time 1416.68, it/s 0.71
  1495. [0130 17:35:08 @multigpu.py:323] [p0115]  step: count(195), step_time 1346.86, mean_step_time 1404.16, it/s 0.71
  1496. [0130 17:35:08 @multigpu.py:323] [p0576]  step: count(219), step_time 1433.47, mean_step_time 1409.33, it/s 0.71
  1497. [0130 17:35:08 @multigpu.py:323] [p0574]  step: count(213), step_time 1362.13, mean_step_time 1415.17, it/s 0.71
  1498. [0130 17:35:09 @multigpu.py:323] [p0115]  step: count(196), step_time 1389.49, mean_step_time 1405.4, it/s 0.71
  1499. [0130 17:35:09 @multigpu.py:323] [p0576]  step: count(220), step_time 1391.26, mean_step_time 1411.76, it/s 0.71
  1500. [0130 17:35:10 @multigpu.py:323] [p0574]  step: count(214), step_time 1386.5, mean_step_time 1414.5, it/s 0.71
  1501. [0130 17:35:11 @multigpu.py:323] [p0576]  step: count(221), step_time 1364.82, mean_step_time 1411.23, it/s 0.71
  1502. [0130 17:35:11 @multigpu.py:323] [p0115]  step: count(197), step_time 1424.18, mean_step_time 1405.53, it/s 0.71
  1503. [0130 17:35:11 @multigpu.py:323] [p0574]  step: count(215), step_time 1408.09, mean_step_time 1411.86, it/s 0.71
  1504. [0130 17:35:12 @multigpu.py:323] [p0576]  step: count(222), step_time 1351.14, mean_step_time 1409.33, it/s 0.71
  1505. [0130 17:35:12 @multigpu.py:323] [p0115]  step: count(198), step_time 1369.54, mean_step_time 1407.69, it/s 0.71
  1506. [0130 17:35:12 @multigpu.py:323] [p0574]  step: count(216), step_time 1432.83, mean_step_time 1414.93, it/s 0.71
  1507. [0130 17:35:14 @multigpu.py:323] [p0576]  step: count(223), step_time 1396.41, mean_step_time 1411.17, it/s 0.71
  1508. [0130 17:35:14 @multigpu.py:323] [p0115]  step: count(199), step_time 1409.2, mean_step_time 1409.23, it/s 0.71
  1509. [0130 17:35:14 @multigpu.py:323] [p0574]  step: count(217), step_time 1347.35, mean_step_time 1408.04, it/s 0.71
  1510. [0130 17:35:15 @multigpu.py:323] [p0115]  step: count(200), step_time 1346.51, mean_step_time 1407.58, it/s 0.71
  1511. sending debugging info...
  1512. sending to address tcp://p0112:61216
  1513. ##### Sending to neptune:  mean_delay :  0.0873081782791 , 0.0 #####
  1514. sending to address tcp://p0112:61216
  1515. ##### Sending to neptune:  max_delay :  0.0873081782791 , -0.0 #####
  1516. ##### Sending to neptune:  min_delay :  0.0873081782791 , -0.0 #####
  1517. [u'delays', [0.0, -0.0, -0.0]]
  1518. ##### Sending to neptune:  active_workers :  0.0873083741135 , 3 #####
  1519. receiving
  1520. ##### Sending to neptune:  cost :  0.0873086949852 , -0.0134393870831 #####
  1521. sending to address tcp://p0112:61216
  1522. ##### Sending to neptune:  policy_loss :  0.0873086949852 , -0.510861754417 #####
  1523. ##### Sending to neptune:  xentropy_loss :  0.0873086949852 , -2.28939032555 #####
  1524. ##### Sending to neptune:  value_loss :  0.0873086949852 , 1.08001065254 #####
  1525. ##### Sending to neptune:  advantage :  0.0873086949852 , 0.00229038670659 #####
  1526. ##### Sending to neptune:  pred_reward :  0.0873086949852 , 0.375006139278 #####
  1527. ##### Sending to neptune:  max_logit :  0.0873086949852 , 0.191750437021 #####
  1528. [u'loss', -0.013439387083053589, -0.5108617544174194, -2.2893903255462646, 1.0800106525421143, 0.0022903867065906525, 0.37500613927841187, 0.1917504370212555]
  1529. receiving
  1530. ##### Sending to neptune:  active_relus :  0.0873091855314 , 8887920.51 #####
  1531. ##### Sending to neptune:  dp_per_s :  0.0873091855314 , 90.4954598592 #####
  1532. [u'other', 8887920.51, 90.49545985924865]
  1533. receiving
  1534. [0130 17:35:15 @multigpu.py:323] [p0576]  step: count(224), step_time 1399.52, mean_step_time 1412.1, it/s 0.71
  1535. [0130 17:35:15 @multigpu.py:323] [p0574]  step: count(218), step_time 1376.75, mean_step_time 1401.25, it/s 0.71
  1536. [0130 17:35:16 @multigpu.py:323] [p0115]  step: count(201), step_time 1388.62, mean_step_time 1409.24, it/s 0.71
  1537. [0130 17:35:16 @multigpu.py:323] [p0576]  step: count(225), step_time 1385.96, mean_step_time 1413.44, it/s 0.71
  1538. [0130 17:35:16 @multigpu.py:323] [p0574]  step: count(219), step_time 1376.69, mean_step_time 1402.03, it/s 0.71
  1539. [0130 17:35:18 @multigpu.py:323] [p0115]  step: count(202), step_time 1362.17, mean_step_time 1407.36, it/s 0.71
  1540. [0130 17:35:18 @multigpu.py:323] [p0576]  step: count(226), step_time 1362.54, mean_step_time 1412.38, it/s 0.71
  1541. [0130 17:35:18 @multigpu.py:323] [p0574]  step: count(220), step_time 1402.37, mean_step_time 1400.41, it/s 0.71
  1542. sending to address tcp://p0112:61216
  1543. ##### Sending to neptune:  online_score :  0.0882514672147 , 1.4 #####
  1544. [u'online', 1.4]
  1545. receiving
  1546. [0130 17:35:19 @multigpu.py:323] [p0576]  step: count(227), step_time 1344.1, mean_step_time 1411.23, it/s 0.71
  1547. [0130 17:35:19 @multigpu.py:323] [p0115]  step: count(203), step_time 1401.89, mean_step_time 1407.59, it/s 0.71
  1548. [0130 17:35:19 @multigpu.py:323] [p0574]  step: count(221), step_time 1395.37, mean_step_time 1399.91, it/s 0.71
  1549. [0130 17:35:20 @multigpu.py:323] [p0576]  step: count(228), step_time 1371.69, mean_step_time 1408.27, it/s 0.71
  1550. [0130 17:35:20 @multigpu.py:323] [p0115]  step: count(204), step_time 1370.23, mean_step_time 1407.16, it/s 0.71
  1551. [0130 17:35:21 @multigpu.py:323] [p0574]  step: count(222), step_time 1453.52, mean_step_time 1402.46, it/s 0.71
  1552. sending to address tcp://p0112:61216
  1553. ##### Sending to neptune:  online_score :  0.0889305941264 , 1.3 #####
  1554. [u'online', 1.3]
  1555. receiving
  1556. [0130 17:35:22 @multigpu.py:323] [p0576]  step: count(229), step_time 1400.69, mean_step_time 1411.7, it/s 0.71
  1557. [0130 17:35:22 @multigpu.py:323] [p0115]  step: count(205), step_time 1375.32, mean_step_time 1406.04, it/s 0.71
  1558. [0130 17:35:22 @multigpu.py:323] [p0574]  step: count(223), step_time 1446.75, mean_step_time 1405.36, it/s 0.71
  1559. [0130 17:35:23 @multigpu.py:323] [p0576]  step: count(230), step_time 1384.77, mean_step_time 1411.95, it/s 0.71
  1560. [0130 17:35:23 @multigpu.py:323] [p0115]  step: count(206), step_time 1368.15, mean_step_time 1406.0, it/s 0.71
  1561. [0130 17:35:24 @multigpu.py:323] [p0574]  step: count(224), step_time 1409.12, mean_step_time 1405.62, it/s 0.71
  1562. [0130 17:35:25 @multigpu.py:323] [p0574]  step: count(225), step_time 1367.4, mean_step_time 1400.45, it/s 0.71
  1563. [0130 17:35:25 @multigpu.py:323] [p0115]  step: count(207), step_time 1781.69, mean_step_time 1400.58, it/s 0.71
  1564. [0130 17:35:25 @multigpu.py:323] [p0576]  step: count(231), step_time 1789.81, mean_step_time 1400.53, it/s 0.71
  1565. [0130 17:35:26 @multigpu.py:323] [p0576]  step: count(232), step_time 1419.14, mean_step_time 1403.28, it/s 0.71
  1566. [0130 17:35:26 @multigpu.py:323] [p0574]  step: count(226), step_time 1427.12, mean_step_time 1400.2, it/s 0.71
  1567. [0130 17:35:26 @multigpu.py:323] [p0115]  step: count(208), step_time 1432.61, mean_step_time 1402.95, it/s 0.71
  1568. [0130 17:35:28 @multigpu.py:323] [p0574]  step: count(227), step_time 1343.76, mean_step_time 1396.1, it/s 0.72
  1569. [0130 17:35:28 @multigpu.py:323] [p0576]  step: count(233), step_time 1364.21, mean_step_time 1402.89, it/s 0.71
  1570. [0130 17:35:28 @multigpu.py:323] [p0115]  step: count(209), step_time 1398.28, mean_step_time 1403.58, it/s 0.71
  1571. sending to address tcp://p0112:61216
  1572. ##### Sending to neptune:  online_score :  0.0910588008165 , 1.4 #####
  1573. [u'online', 1.4]
  1574. receiving
  1575. [0130 17:35:29 @multigpu.py:323] [p0574]  step: count(228), step_time 1385.26, mean_step_time 1397.12, it/s 0.72
  1576. [0130 17:35:29 @multigpu.py:323] [p0576]  step: count(234), step_time 1381.8, mean_step_time 1402.59, it/s 0.71
  1577. [0130 17:35:29 @multigpu.py:323] [p0115]  step: count(210), step_time 1369.82, mean_step_time 1404.83, it/s 0.71
  1578. [0130 17:35:30 @multigpu.py:323] [p0574]  step: count(229), step_time 1344.25, mean_step_time 1393.68, it/s 0.72
  1579. [0130 17:35:31 @multigpu.py:323] [p0576]  step: count(235), step_time 1426.86, mean_step_time 1406.13, it/s 0.71
  1580. [0130 17:35:31 @multigpu.py:323] [p0115]  step: count(211), step_time 1409.13, mean_step_time 1405.16, it/s 0.71
  1581. [0130 17:35:32 @multigpu.py:323] [p0574]  step: count(230), step_time 1422.48, mean_step_time 1394.77, it/s 0.72
  1582. [0130 17:35:32 @multigpu.py:323] [p0115]  step: count(212), step_time 1342.63, mean_step_time 1404.12, it/s 0.71
  1583. [0130 17:35:32 @multigpu.py:323] [p0576]  step: count(236), step_time 1360.8, mean_step_time 1404.35, it/s 0.71
  1584. [0130 17:35:33 @multigpu.py:323] [p0574]  step: count(231), step_time 1366.62, mean_step_time 1392.67, it/s 0.72
  1585. [0130 17:35:33 @multigpu.py:323] [p0576]  step: count(237), step_time 1404.87, mean_step_time 1405.15, it/s 0.71
  1586. [0130 17:35:33 @multigpu.py:323] [p0115]  step: count(213), step_time 1440.57, mean_step_time 1406.55, it/s 0.71
  1587. [0130 17:35:35 @multigpu.py:323] [p0576]  step: count(238), step_time 1372.38, mean_step_time 1405.31, it/s 0.71
  1588. [0130 17:35:35 @multigpu.py:323] [p0574]  step: count(232), step_time 1442.58, mean_step_time 1394.85, it/s 0.72
  1589. [0130 17:35:35 @multigpu.py:323] [p0115]  step: count(214), step_time 1366.34, mean_step_time 1404.66, it/s 0.71
  1590. [0130 17:35:36 @multigpu.py:323] [p0576]  step: count(239), step_time 1396.97, mean_step_time 1403.49, it/s 0.71
  1591. [0130 17:35:36 @multigpu.py:323] [p0115]  step: count(215), step_time 1435.04, mean_step_time 1409.07, it/s 0.71
  1592. [0130 17:35:36 @multigpu.py:323] [p0574]  step: count(233), step_time 1562.81, mean_step_time 1404.88, it/s 0.71
  1593. [0130 17:35:37 @multigpu.py:323] [p0576]  step: count(240), step_time 1401.66, mean_step_time 1404.01, it/s 0.71
  1594. [0130 17:35:38 @multigpu.py:323] [p0115]  step: count(216), step_time 1426.44, mean_step_time 1410.92, it/s 0.71
  1595. [0130 17:35:38 @multigpu.py:323] [p0574]  step: count(234), step_time 1530.0, mean_step_time 1412.06, it/s 0.71
  1596. [0130 17:35:39 @multigpu.py:323] [p0576]  step: count(241), step_time 1389.36, mean_step_time 1405.23, it/s 0.71
  1597. [0130 17:35:39 @multigpu.py:323] [p0115]  step: count(217), step_time 1348.05, mean_step_time 1407.11, it/s 0.71
  1598. [0130 17:35:39 @multigpu.py:323] [p0574]  step: count(235), step_time 1505.13, mean_step_time 1416.91, it/s 0.71
  1599. [0130 17:35:40 @multigpu.py:323] [p0576]  step: count(242), step_time 1439.39, mean_step_time 1409.65, it/s 0.71
  1600. [0130 17:35:40 @multigpu.py:323] [p0115]  step: count(218), step_time 1401.83, mean_step_time 1408.73, it/s 0.71
  1601. [0130 17:35:41 @multigpu.py:323] [p0574]  step: count(236), step_time 1410.48, mean_step_time 1415.79, it/s 0.71
  1602. [0130 17:35:42 @multigpu.py:323] [p0576]  step: count(243), step_time 1347.01, mean_step_time 1407.18, it/s 0.71
  1603. [0130 17:35:42 @multigpu.py:323] [p0115]  step: count(219), step_time 1380.97, mean_step_time 1407.31, it/s 0.71
  1604. [0130 17:35:42 @multigpu.py:323] [p0574]  step: count(237), step_time 1423.83, mean_step_time 1419.62, it/s 0.7
  1605. [0130 17:35:43 @multigpu.py:323] [p0576]  step: count(244), step_time 1387.0, mean_step_time 1406.55, it/s 0.71
  1606. [0130 17:35:43 @multigpu.py:323] [p0115]  step: count(220), step_time 1371.53, mean_step_time 1408.57, it/s 0.71
  1607. [0130 17:35:44 @multigpu.py:323] [p0574]  step: count(238), step_time 1374.06, mean_step_time 1419.48, it/s 0.7
  1608. [0130 17:35:44 @multigpu.py:323] [p0576]  step: count(245), step_time 1384.53, mean_step_time 1406.48, it/s 0.71
  1609. [0130 17:35:44 @multigpu.py:323] [p0115]  step: count(221), step_time 1386.56, mean_step_time 1408.46, it/s 0.71
  1610. [0130 17:35:45 @multigpu.py:323] [p0574]  step: count(239), step_time 1363.55, mean_step_time 1418.82, it/s 0.7
  1611. [0130 17:35:46 @multigpu.py:323] [p0576]  step: count(246), step_time 1337.92, mean_step_time 1405.25, it/s 0.71
  1612. [0130 17:35:46 @multigpu.py:323] [p0115]  step: count(222), step_time 1363.66, mean_step_time 1408.54, it/s 0.71
  1613. [0130 17:35:46 @multigpu.py:323] [p0574]  step: count(240), step_time 1375.6, mean_step_time 1417.48, it/s 0.71
  1614. [0130 17:35:47 @multigpu.py:323] [p0576]  step: count(247), step_time 1425.29, mean_step_time 1409.31, it/s 0.71
  1615. [0130 17:35:47 @multigpu.py:323] [p0115]  step: count(223), step_time 1375.84, mean_step_time 1407.23, it/s 0.71
  1616. [0130 17:35:48 @multigpu.py:323] [p0574]  step: count(241), step_time 1392.03, mean_step_time 1417.32, it/s 0.71
  1617. sending to address tcp://p0112:61216
  1618. ##### Sending to neptune:  online_score :  0.0964728083213 , 0.7 #####
  1619. [u'online', 0.7]
  1620. receiving
  1621. sending to address tcp://p0112:61216
  1622. ##### Sending to neptune:  online_score :  0.0965966761112 , 2.3 #####
  1623. [u'online', 2.3]
  1624. receiving
  1625. [0130 17:35:49 @multigpu.py:323] [p0576]  step: count(248), step_time 1370.39, mean_step_time 1409.24, it/s 0.71
  1626. [0130 17:35:49 @multigpu.py:323] [p0115]  step: count(224), step_time 1380.98, mean_step_time 1407.77, it/s 0.71
  1627. [0130 17:35:49 @multigpu.py:323] [p0574]  step: count(242), step_time 1367.97, mean_step_time 1413.04, it/s 0.71
  1628. [0130 17:35:50 @multigpu.py:323] [p0576]  step: count(249), step_time 1371.62, mean_step_time 1407.79, it/s 0.71
  1629. [0130 17:35:50 @multigpu.py:323] [p0115]  step: count(225), step_time 1399.34, mean_step_time 1408.97, it/s 0.71
  1630. [0130 17:35:50 @multigpu.py:323] [p0574]  step: count(243), step_time 1377.29, mean_step_time 1409.57, it/s 0.71
  1631. [0130 17:35:51 @multigpu.py:323] [p0576]  step: count(250), step_time 1401.31, mean_step_time 1408.62, it/s 0.71
  1632. [0130 17:35:51 @multigpu.py:323] [p0115]  step: count(226), step_time 1427.54, mean_step_time 1411.94, it/s 0.71
  1633. [0130 17:35:52 @multigpu.py:323] [p0574]  step: count(244), step_time 1402.78, mean_step_time 1409.25, it/s 0.71
  1634. [0130 17:35:53 @multigpu.py:323] [p0574]  step: count(245), step_time 1371.89, mean_step_time 1409.47, it/s 0.71
  1635. [0130 17:35:53 @multigpu.py:323] [p0576]  step: count(251), step_time 1805.83, mean_step_time 1409.42, it/s 0.71
  1636. [0130 17:35:53 @multigpu.py:323] [p0115]  step: count(227), step_time 1732.07, mean_step_time 1409.46, it/s 0.71
  1637. sending to address tcp://p0112:61216
  1638. ##### Sending to neptune:  online_score :  0.0980394785934 , 1.9 #####
  1639. [u'online', 1.9]
  1640. receiving
  1641. [0130 17:35:55 @multigpu.py:323] [p0115]  step: count(228), step_time 1372.59, mean_step_time 1406.46, it/s 0.71
  1642. [0130 17:35:55 @multigpu.py:323] [p0574]  step: count(246), step_time 1373.57, mean_step_time 1406.8, it/s 0.71
  1643. [0130 17:35:55 @multigpu.py:323] [p0576]  step: count(252), step_time 1411.17, mean_step_time 1409.02, it/s 0.71
  1644. [0130 17:35:56 @multigpu.py:323] [p0115]  step: count(229), step_time 1372.77, mean_step_time 1405.18, it/s 0.71
  1645. [0130 17:35:56 @multigpu.py:323] [p0574]  step: count(247), step_time 1372.83, mean_step_time 1408.25, it/s 0.71
  1646. [0130 17:35:56 @multigpu.py:323] [p0576]  step: count(253), step_time 1386.74, mean_step_time 1410.14, it/s 0.71
  1647. [0130 17:35:57 @multigpu.py:323] [p0115]  step: count(230), step_time 1386.84, mean_step_time 1406.04, it/s 0.71
  1648. [0130 17:35:57 @multigpu.py:323] [p0574]  step: count(248), step_time 1412.52, mean_step_time 1409.61, it/s 0.71
  1649. [0130 17:35:57 @multigpu.py:323] [p0576]  step: count(254), step_time 1386.09, mean_step_time 1410.36, it/s 0.71
  1650. [0130 17:35:59 @multigpu.py:323] [p0574]  step: count(249), step_time 1337.01, mean_step_time 1409.25, it/s 0.71
  1651. [0130 17:35:59 @multigpu.py:323] [p0115]  step: count(231), step_time 1365.18, mean_step_time 1403.84, it/s 0.71
  1652. [0130 17:35:59 @multigpu.py:323] [p0576]  step: count(255), step_time 1402.17, mean_step_time 1409.13, it/s 0.71
  1653. [0130 17:36:00 @multigpu.py:323] [p0115]  step: count(232), step_time 1371.18, mean_step_time 1405.27, it/s 0.71
  1654. [0130 17:36:00 @multigpu.py:323] [p0574]  step: count(250), step_time 1394.7, mean_step_time 1407.86, it/s 0.71
  1655. [0130 17:36:00 @multigpu.py:323] [p0576]  step: count(256), step_time 1370.45, mean_step_time 1409.61, it/s 0.71
  1656. [0130 17:36:01 @multigpu.py:323] [p0115]  step: count(233), step_time 1407.6, mean_step_time 1403.62, it/s 0.71
  1657. [0130 17:36:01 @multigpu.py:323] [p0574]  step: count(251), step_time 1392.27, mean_step_time 1409.14, it/s 0.71
  1658. [0130 17:36:02 @multigpu.py:323] [p0576]  step: count(257), step_time 1404.48, mean_step_time 1409.59, it/s 0.71
  1659. [0130 17:36:03 @multigpu.py:323] [p0115]  step: count(234), step_time 1388.9, mean_step_time 1404.75, it/s 0.71
  1660. [0130 17:36:03 @multigpu.py:323] [p0574]  step: count(252), step_time 1392.01, mean_step_time 1406.62, it/s 0.71
  1661. [0130 17:36:03 @multigpu.py:323] [p0576]  step: count(258), step_time 1382.35, mean_step_time 1410.09, it/s 0.71
  1662. [0130 17:36:04 @multigpu.py:323] [p0574]  step: count(253), step_time 1358.53, mean_step_time 1396.4, it/s 0.72
  1663. [0130 17:36:04 @multigpu.py:323] [p0115]  step: count(235), step_time 1397.92, mean_step_time 1402.89, it/s 0.71
  1664. [0130 17:36:04 @multigpu.py:323] [p0576]  step: count(259), step_time 1351.43, mean_step_time 1407.81, it/s 0.71
  1665. [0130 17:36:06 @multigpu.py:323] [p0115]  step: count(236), step_time 1348.7, mean_step_time 1399.0, it/s 0.71
  1666. [0130 17:36:06 @multigpu.py:323] [p0576]  step: count(260), step_time 1386.78, mean_step_time 1407.07, it/s 0.71
  1667. [0130 17:36:06 @multigpu.py:323] [p0574]  step: count(254), step_time 1552.29, mean_step_time 1397.52, it/s 0.72
  1668. [0130 17:36:07 @multigpu.py:323] [p0115]  step: count(237), step_time 1384.47, mean_step_time 1400.82, it/s 0.71
  1669. [0130 17:36:07 @multigpu.py:323] [p0576]  step: count(261), step_time 1396.65, mean_step_time 1407.43, it/s 0.71
  1670. [0130 17:36:07 @multigpu.py:323] [p0574]  step: count(255), step_time 1380.3, mean_step_time 1391.28, it/s 0.72
  1671. [0130 17:36:08 @multigpu.py:323] [p0115]  step: count(238), step_time 1420.84, mean_step_time 1401.77, it/s 0.71
  1672. [0130 17:36:08 @multigpu.py:323] [p0576]  step: count(262), step_time 1414.32, mean_step_time 1406.18, it/s 0.71
  1673. [0130 17:36:09 @multigpu.py:323] [p0574]  step: count(256), step_time 1380.86, mean_step_time 1389.79, it/s 0.72
  1674. [0130 17:36:10 @multigpu.py:323] [p0115]  step: count(239), step_time 1345.4, mean_step_time 1400.0, it/s 0.71
  1675. [0130 17:36:10 @multigpu.py:323] [p0576]  step: count(263), step_time 1396.73, mean_step_time 1408.66, it/s 0.71
  1676. [0130 17:36:10 @multigpu.py:323] [p0574]  step: count(257), step_time 1388.26, mean_step_time 1388.02, it/s 0.72
  1677. sending to address tcp://p0112:61216
  1678. ##### Sending to neptune:  online_score :  0.102663866348 , 1.3 #####
  1679. [u'online', 1.3]
  1680. receiving
  1681. [0130 17:36:11 @multigpu.py:323] [p0115]  step: count(240), step_time 1385.78, mean_step_time 1400.71, it/s 0.71
  1682. [0130 17:36:11 @multigpu.py:323] [p0576]  step: count(264), step_time 1350.94, mean_step_time 1406.86, it/s 0.71
  1683. [0130 17:36:11 @multigpu.py:323] [p0574]  step: count(258), step_time 1439.78, mean_step_time 1391.3, it/s 0.72
  1684. sending to address tcp://p0112:61216
  1685. ##### Sending to neptune:  online_score :  0.103130486078 , 1.1 #####
  1686. [u'online', 1.1]
  1687. receiving
  1688. [0130 17:36:12 @multigpu.py:323] [p0115]  step: count(241), step_time 1377.61, mean_step_time 1400.26, it/s 0.71
  1689. [0130 17:36:13 @multigpu.py:323] [p0576]  step: count(265), step_time 1359.71, mean_step_time 1405.62, it/s 0.71
  1690. [0130 17:36:13 @multigpu.py:323] [p0574]  step: count(259), step_time 1362.89, mean_step_time 1391.27, it/s 0.72
  1691. [0130 17:36:14 @multigpu.py:323] [p0115]  step: count(242), step_time 1394.91, mean_step_time 1401.82, it/s 0.71
  1692. [0130 17:36:14 @multigpu.py:323] [p0576]  step: count(266), step_time 1413.71, mean_step_time 1409.41, it/s 0.71
  1693. [0130 17:36:14 @multigpu.py:323] [p0574]  step: count(260), step_time 1389.64, mean_step_time 1391.97, it/s 0.72
  1694. [0130 17:36:15 @multigpu.py:323] [p0115]  step: count(243), step_time 1373.49, mean_step_time 1401.7, it/s 0.71
  1695. [0130 17:36:15 @multigpu.py:323] [p0576]  step: count(267), step_time 1374.35, mean_step_time 1406.86, it/s 0.71
  1696. [0130 17:36:15 @multigpu.py:323] [p0574]  step: count(261), step_time 1348.26, mean_step_time 1389.78, it/s 0.72
  1697. [0130 17:36:17 @multigpu.py:323] [p0115]  step: count(244), step_time 1386.47, mean_step_time 1401.98, it/s 0.71
  1698. [0130 17:36:17 @multigpu.py:323] [p0576]  step: count(268), step_time 1384.0, mean_step_time 1407.54, it/s 0.71
  1699. [0130 17:36:17 @multigpu.py:323] [p0574]  step: count(262), step_time 1349.67, mean_step_time 1388.87, it/s 0.72
  1700. [0130 17:36:18 @multigpu.py:323] [p0115]  step: count(245), step_time 1383.07, mean_step_time 1401.17, it/s 0.71
  1701. [0130 17:36:18 @multigpu.py:323] [p0576]  step: count(269), step_time 1380.73, mean_step_time 1408.0, it/s 0.71
  1702. [0130 17:36:18 @multigpu.py:323] [p0574]  step: count(263), step_time 1389.54, mean_step_time 1389.48, it/s 0.72
  1703. [0130 17:36:19 @multigpu.py:323] [p0115]  step: count(246), step_time 1361.13, mean_step_time 1397.85, it/s 0.72
  1704. [0130 17:36:20 @multigpu.py:323] [p0574]  step: count(264), step_time 1324.5, mean_step_time 1385.57, it/s 0.72
  1705. [0130 17:36:20 @multigpu.py:323] [p0576]  step: count(270), step_time 1399.82, mean_step_time 1407.92, it/s 0.71
  1706. sending to address tcp://p0112:61216
  1707. ##### Sending to neptune:  online_score :  0.10532262471 , 0.9 #####
  1708. [u'online', 0.9]
  1709. ##### Sending to neptune:  active_workers :  0.105322658552 , 3 #####
  1710. receiving
  1711. [0130 17:36:21 @multigpu.py:323] [p0574]  step: count(265), step_time 1493.54, mean_step_time 1391.65, it/s 0.72
  1712. [0130 17:36:21 @multigpu.py:323] [p0576]  step: count(271), step_time 1479.04, mean_step_time 1391.58, it/s 0.72
  1713. [0130 17:36:21 @multigpu.py:323] [p0115]  step: count(247), step_time 1609.22, mean_step_time 1391.7, it/s 0.72
  1714. [0130 17:36:22 @multigpu.py:323] [p0115]  step: count(248), step_time 1364.46, mean_step_time 1391.3, it/s 0.72
  1715. [0130 17:36:22 @multigpu.py:323] [p0576]  step: count(272), step_time 1373.2, mean_step_time 1389.69, it/s 0.72
  1716. [0130 17:36:23 @multigpu.py:323] [p0574]  step: count(266), step_time 1546.98, mean_step_time 1400.32, it/s 0.71
  1717. [0130 17:36:24 @multigpu.py:323] [p0115]  step: count(249), step_time 1400.73, mean_step_time 1392.69, it/s 0.72
  1718. [0130 17:36:24 @multigpu.py:323] [p0576]  step: count(273), step_time 1402.29, mean_step_time 1390.46, it/s 0.72
  1719. [0130 17:36:24 @multigpu.py:323] [p0574]  step: count(267), step_time 1500.75, mean_step_time 1406.72, it/s 0.71
  1720. [0130 17:36:25 @multigpu.py:323] [p0115]  step: count(250), step_time 1340.03, mean_step_time 1390.35, it/s 0.72
  1721. [0130 17:36:25 @multigpu.py:323] [p0576]  step: count(274), step_time 1363.4, mean_step_time 1389.33, it/s 0.72
  1722. [0130 17:36:26 @multigpu.py:323] [p0574]  step: count(268), step_time 1638.14, mean_step_time 1418.0, it/s 0.71
  1723. [0130 17:36:27 @multigpu.py:323] [p0115]  step: count(251), step_time 1383.45, mean_step_time 1391.27, it/s 0.72
  1724. [0130 17:36:27 @multigpu.py:323] [p0576]  step: count(275), step_time 1434.35, mean_step_time 1390.94, it/s 0.72
  1725. [0130 17:36:27 @multigpu.py:323] [p0574]  step: count(269), step_time 1532.01, mean_step_time 1427.75, it/s 0.7
  1726. [0130 17:36:28 @multigpu.py:323] [p0115]  step: count(252), step_time 1385.99, mean_step_time 1392.01, it/s 0.72
  1727. [0130 17:36:28 @multigpu.py:323] [p0576]  step: count(276), step_time 1401.97, mean_step_time 1392.51, it/s 0.72
  1728. [0130 17:36:29 @multigpu.py:323] [p0574]  step: count(270), step_time 1592.14, mean_step_time 1437.62, it/s 0.7
  1729. [0130 17:36:29 @multigpu.py:323] [p0115]  step: count(253), step_time 1385.17, mean_step_time 1390.89, it/s 0.72
  1730. [0130 17:36:29 @multigpu.py:323] [p0576]  step: count(277), step_time 1378.67, mean_step_time 1391.22, it/s 0.72
  1731. [0130 17:36:30 @multigpu.py:323] [p0574]  step: count(271), step_time 1486.68, mean_step_time 1442.34, it/s 0.69
  1732. [0130 17:36:31 @multigpu.py:323] [p0115]  step: count(254), step_time 1391.45, mean_step_time 1391.01, it/s 0.72
  1733. [0130 17:36:31 @multigpu.py:323] [p0576]  step: count(278), step_time 1465.1, mean_step_time 1395.36, it/s 0.72
  1734. [0130 17:36:32 @multigpu.py:323] [p0574]  step: count(272), step_time 1466.54, mean_step_time 1446.07, it/s 0.69
  1735. [0130 17:36:32 @multigpu.py:323] [p0115]  step: count(255), step_time 1385.32, mean_step_time 1390.38, it/s 0.72
  1736. [0130 17:36:32 @multigpu.py:323] [p0576]  step: count(279), step_time 1373.63, mean_step_time 1396.47, it/s 0.72
  1737. sending to address tcp://p0112:61216
  1738. ##### Sending to neptune:  online_score :  0.108991423051 , 1.3 #####
  1739. [u'online', 1.3]
  1740. receiving
  1741. [0130 17:36:33 @multigpu.py:323] [p0574]  step: count(273), step_time 1636.7, mean_step_time 1459.97, it/s 0.68
  1742. [0130 17:36:33 @multigpu.py:323] [p0115]  step: count(256), step_time 1384.76, mean_step_time 1392.19, it/s 0.72
  1743. [0130 17:36:34 @multigpu.py:323] [p0576]  step: count(280), step_time 1397.0, mean_step_time 1396.98, it/s 0.72
  1744. [0130 17:36:35 @multigpu.py:323] [p0115]  step: count(257), step_time 1385.74, mean_step_time 1392.25, it/s 0.72
  1745. [0130 17:36:35 @multigpu.py:323] [p0576]  step: count(281), step_time 1343.67, mean_step_time 1394.33, it/s 0.72
  1746. [0130 17:36:35 @multigpu.py:323] [p0574]  step: count(274), step_time 1662.91, mean_step_time 1465.5, it/s 0.68
  1747. [0130 17:36:36 @multigpu.py:323] [p0115]  step: count(258), step_time 1358.49, mean_step_time 1389.13, it/s 0.72
  1748. [0130 17:36:36 @multigpu.py:323] [p0576]  step: count(282), step_time 1396.08, mean_step_time 1393.42, it/s 0.72
  1749. [0130 17:36:37 @multigpu.py:323] [p0574]  step: count(275), step_time 1619.75, mean_step_time 1477.48, it/s 0.68
  1750. [0130 17:36:38 @multigpu.py:323] [p0115]  step: count(259), step_time 1387.44, mean_step_time 1391.24, it/s 0.72
  1751. [0130 17:36:38 @multigpu.py:323] [p0576]  step: count(283), step_time 1457.86, mean_step_time 1396.48, it/s 0.72
  1752. [0130 17:36:38 @multigpu.py:323] [p0574]  step: count(276), step_time 1383.84, mean_step_time 1477.63, it/s 0.68
  1753. [0130 17:36:39 @multigpu.py:323] [p0115]  step: count(260), step_time 1375.11, mean_step_time 1390.7, it/s 0.72
  1754. [0130 17:36:39 @multigpu.py:323] [p0576]  step: count(284), step_time 1445.71, mean_step_time 1401.21, it/s 0.71
  1755. [0130 17:36:39 @multigpu.py:323] [p0574]  step: count(277), step_time 1414.63, mean_step_time 1478.94, it/s 0.68
  1756. sending to address tcp://p0112:61216
  1757. ##### Sending to neptune:  online_score :  0.110936076641 , 1.3 #####
  1758. [u'online', 1.3]
  1759. receiving
  1760. [0130 17:36:40 @multigpu.py:323] [p0115]  step: count(261), step_time 1376.47, mean_step_time 1390.64, it/s 0.72
  1761. [0130 17:36:41 @multigpu.py:323] [p0576]  step: count(285), step_time 1379.43, mean_step_time 1402.2, it/s 0.71
  1762. [0130 17:36:41 @multigpu.py:323] [p0574]  step: count(278), step_time 1411.53, mean_step_time 1477.53, it/s 0.68
  1763. [0130 17:36:42 @multigpu.py:323] [p0115]  step: count(262), step_time 1371.64, mean_step_time 1389.48, it/s 0.72
  1764. [0130 17:36:42 @multigpu.py:323] [p0576]  step: count(286), step_time 1378.01, mean_step_time 1400.42, it/s 0.71
  1765. [0130 17:36:42 @multigpu.py:323] [p0574]  step: count(279), step_time 1407.81, mean_step_time 1479.78, it/s 0.68
  1766. [0130 17:36:43 @multigpu.py:323] [p0115]  step: count(263), step_time 1485.85, mean_step_time 1395.1, it/s 0.72
  1767. [0130 17:36:43 @multigpu.py:323] [p0576]  step: count(287), step_time 1401.66, mean_step_time 1401.78, it/s 0.71
  1768. [0130 17:36:44 @multigpu.py:323] [p0574]  step: count(280), step_time 1467.43, mean_step_time 1483.67, it/s 0.67
  1769. [0130 17:36:45 @multigpu.py:323] [p0115]  step: count(264), step_time 1356.22, mean_step_time 1393.59, it/s 0.72
  1770. sending to address tcp://p0112:61216
  1771. ##### Sending to neptune:  online_score :  0.112287428313 , 1.3 #####
  1772. [u'online', 1.3]
  1773. receiving
  1774. [0130 17:36:45 @multigpu.py:323] [p0576]  step: count(288), step_time 1476.2, mean_step_time 1406.39, it/s 0.71
  1775. [0130 17:36:45 @multigpu.py:323] [p0574]  step: count(281), step_time 1594.94, mean_step_time 1496.0, it/s 0.67
  1776. [0130 17:36:46 @multigpu.py:323] [p0115]  step: count(265), step_time 1363.91, mean_step_time 1392.63, it/s 0.72
  1777. [0130 17:36:46 @multigpu.py:323] [p0576]  step: count(289), step_time 1398.59, mean_step_time 1407.28, it/s 0.71
  1778. [0130 17:36:47 @multigpu.py:323] [p0574]  step: count(282), step_time 1580.49, mean_step_time 1507.54, it/s 0.66
  1779. [0130 17:36:47 @multigpu.py:323] [p0115]  step: count(266), step_time 1397.36, mean_step_time 1394.44, it/s 0.72
  1780. [0130 17:36:48 @multigpu.py:323] [p0576]  step: count(290), step_time 1325.89, mean_step_time 1403.59, it/s 0.71
  1781. [0130 17:36:48 @multigpu.py:323] [p0574]  step: count(283), step_time 1467.9, mean_step_time 1511.46, it/s 0.66
  1782. [0130 17:36:49 @multigpu.py:323] [p0115]  step: count(267), step_time 1365.89, mean_step_time 1382.27, it/s 0.72
  1783. [0130 17:36:50 @multigpu.py:323] [p0115]  step: count(268), step_time 1384.45, mean_step_time 1383.27, it/s 0.72
  1784. [0130 17:36:50 @multigpu.py:323] [p0574]  step: count(284), step_time 1619.19, mean_step_time 1526.2, it/s 0.66
  1785. [0130 17:36:50 @multigpu.py:323] [p0576]  step: count(291), step_time 2436.1, mean_step_time 1451.44, it/s 0.69
  1786. [0130 17:36:51 @multigpu.py:323] [p0576]  step: count(292), step_time 1357.52, mean_step_time 1450.66, it/s 0.69
  1787. [0130 17:36:51 @multigpu.py:323] [p0115]  step: count(269), step_time 1398.63, mean_step_time 1383.17, it/s 0.72
  1788. [0130 17:36:52 @multigpu.py:323] [p0574]  step: count(285), step_time 1613.86, mean_step_time 1532.21, it/s 0.65
  1789. [0130 17:36:53 @multigpu.py:323] [p0115]  step: count(270), step_time 1342.82, mean_step_time 1383.31, it/s 0.72
  1790. [0130 17:36:53 @multigpu.py:323] [p0576]  step: count(293), step_time 1395.05, mean_step_time 1450.29, it/s 0.69
  1791. [0130 17:36:53 @multigpu.py:323] [p0574]  step: count(286), step_time 1630.58, mean_step_time 1536.39, it/s 0.65
  1792. [0130 17:36:54 @multigpu.py:323] [p0115]  step: count(271), step_time 1308.8, mean_step_time 1379.58, it/s 0.72
  1793. [0130 17:36:54 @multigpu.py:323] [p0576]  step: count(294), step_time 1346.21, mean_step_time 1449.44, it/s 0.69
  1794. [0130 17:36:55 @multigpu.py:323] [p0574]  step: count(287), step_time 1521.75, mean_step_time 1537.44, it/s 0.65
  1795. [0130 17:36:56 @multigpu.py:323] [p0115]  step: count(272), step_time 1412.48, mean_step_time 1380.9, it/s 0.72
  1796. [0130 17:36:56 @multigpu.py:323] [p0576]  step: count(295), step_time 1369.98, mean_step_time 1446.22, it/s 0.69
  1797. [0130 17:36:56 @multigpu.py:323] [p0574]  step: count(288), step_time 1358.74, mean_step_time 1523.47, it/s 0.66
  1798. [0130 17:36:57 @multigpu.py:323] [p0115]  step: count(273), step_time 1386.04, mean_step_time 1380.94, it/s 0.72
  1799. [0130 17:36:57 @multigpu.py:323] [p0576]  step: count(296), step_time 1424.16, mean_step_time 1447.33, it/s 0.69
  1800. [0130 17:36:58 @multigpu.py:323] [p0574]  step: count(289), step_time 1388.73, mean_step_time 1516.31, it/s 0.66
  1801. sending to address tcp://p0112:61216
  1802. ##### Sending to neptune:  online_score :  0.115880426632 , 1.2 #####
  1803. [u'online', 1.2]
  1804. receiving
  1805. [0130 17:36:58 @multigpu.py:323] [p0115]  step: count(274), step_time 1380.34, mean_step_time 1380.39, it/s 0.72
  1806. [0130 17:36:58 @multigpu.py:323] [p0576]  step: count(297), step_time 1368.9, mean_step_time 1446.84, it/s 0.69
  1807. [0130 17:36:59 @multigpu.py:323] [p0574]  step: count(290), step_time 1394.5, mean_step_time 1506.43, it/s 0.66
  1808. [0130 17:37:00 @multigpu.py:323] [p0115]  step: count(275), step_time 1379.69, mean_step_time 1380.11, it/s 0.72
  1809. [0130 17:37:00 @multigpu.py:323] [p0576]  step: count(298), step_time 1394.81, mean_step_time 1443.32, it/s 0.69
  1810. [0130 17:37:00 @multigpu.py:323] [p0574]  step: count(291), step_time 1387.45, mean_step_time 1501.46, it/s 0.67
  1811. [0130 17:37:01 @multigpu.py:323] [p0115]  step: count(276), step_time 1387.1, mean_step_time 1380.22, it/s 0.72
  1812. [0130 17:37:01 @multigpu.py:323] [p0576]  step: count(299), step_time 1385.73, mean_step_time 1443.93, it/s 0.69
  1813. [0130 17:37:02 @multigpu.py:323] [p0574]  step: count(292), step_time 1369.93, mean_step_time 1496.63, it/s 0.67
  1814. [0130 17:37:02 @multigpu.py:323] [p0115]  step: count(277), step_time 1383.91, mean_step_time 1380.13, it/s 0.72
  1815. [0130 17:37:03 @multigpu.py:323] [p0576]  step: count(300), step_time 1462.3, mean_step_time 1447.19, it/s 0.69
  1816. sending debugging info...
  1817. sending to address tcp://p0112:61216
  1818. ##### Sending to neptune:  mean_delay :  0.117210962176 , 0.0 #####
  1819. sending to address tcp://p0112:61216
  1820. ##### Sending to neptune:  max_delay :  0.117210962176 , -0.0 #####
  1821. ##### Sending to neptune:  min_delay :  0.117210962176 , -0.0 #####
  1822. [u'delays', [0.0, -0.0, -0.0]]
  1823. receiving
  1824. ##### Sending to neptune:  cost :  0.117211548554 , -0.0115024279803 #####
  1825. sending to address tcp://p0112:61216
  1826. ##### Sending to neptune:  policy_loss :  0.117211548554 , -0.317066699266 #####
  1827. ##### Sending to neptune:  xentropy_loss :  0.117211548554 , -2.28820109367 #####
  1828. ##### Sending to neptune:  value_loss :  0.117211548554 , 1.13295674324 #####
  1829. ##### Sending to neptune:  advantage :  0.117211548554 , 0.00142617744859 #####
  1830. ##### Sending to neptune:  pred_reward :  0.117211548554 , 0.396612882614 #####
  1831. ##### Sending to neptune:  max_logit :  0.117211548554 , 0.193715110421 #####
  1832. [u'loss', -0.011502427980303764, -0.3170666992664337, -2.288201093673706, 1.1329567432403564, 0.00142617744859308, 0.39661288261413574, 0.19371511042118073]
  1833. receiving
  1834. ##### Sending to neptune:  active_relus :  0.117212024993 , 8889058.0 #####
  1835. ##### Sending to neptune:  dp_per_s :  0.117212024993 , 91.0211410016 #####
  1836. [u'other', 8889058.0, 91.02114100159582]
  1837. receiving
  1838. [0130 17:37:03 @multigpu.py:323] [p0574]  step: count(293), step_time 1378.74, mean_step_time 1483.74, it/s 0.67
  1839. [0130 17:37:04 @multigpu.py:323] [p0115]  step: count(278), step_time 1361.57, mean_step_time 1380.29, it/s 0.72
  1840. sending to address tcp://p0112:61216
  1841. ##### Sending to neptune:  online_score :  0.117604462769 , 1.4 #####
  1842. [u'online', 1.4]
  1843. receiving
  1844. [0130 17:37:04 @multigpu.py:323] [p0576]  step: count(301), step_time 1422.54, mean_step_time 1451.14, it/s 0.69
  1845. [0130 17:37:05 @multigpu.py:323] [p0574]  step: count(294), step_time 1524.22, mean_step_time 1476.8, it/s 0.68
  1846. [0130 17:37:05 @multigpu.py:323] [p0115]  step: count(279), step_time 1390.55, mean_step_time 1380.44, it/s 0.72
  1847. [0130 17:37:05 @multigpu.py:323] [p0576]  step: count(302), step_time 1373.9, mean_step_time 1450.03, it/s 0.69
  1848. [0130 17:37:06 @multigpu.py:323] [p0574]  step: count(295), step_time 1396.08, mean_step_time 1465.62, it/s 0.68
  1849. [0130 17:37:07 @multigpu.py:323] [p0115]  step: count(280), step_time 1366.44, mean_step_time 1380.01, it/s 0.72
  1850. [0130 17:37:07 @multigpu.py:323] [p0576]  step: count(303), step_time 1444.09, mean_step_time 1449.34, it/s 0.69
  1851. sending to address tcp://p0112:61216
  1852. ##### Sending to neptune:  online_score :  0.118510236078 , 1.2 #####
  1853. [u'online', 1.2]
  1854. receiving
  1855. [0130 17:37:08 @multigpu.py:323] [p0574]  step: count(296), step_time 1630.41, mean_step_time 1477.95, it/s 0.68
  1856. [0130 17:37:08 @multigpu.py:323] [p0115]  step: count(281), step_time 1372.06, mean_step_time 1379.79, it/s 0.72
  1857. [0130 17:37:08 @multigpu.py:323] [p0576]  step: count(304), step_time 1350.36, mean_step_time 1444.57, it/s 0.69
  1858. [0130 17:37:09 @multigpu.py:323] [p0574]  step: count(297), step_time 1582.59, mean_step_time 1486.34, it/s 0.67
  1859. [0130 17:37:09 @multigpu.py:323] [p0115]  step: count(282), step_time 1378.39, mean_step_time 1380.13, it/s 0.72
  1860. [0130 17:37:10 @multigpu.py:323] [p0576]  step: count(305), step_time 1426.73, mean_step_time 1446.94, it/s 0.69
  1861. [0130 17:37:11 @multigpu.py:323] [p0574]  step: count(298), step_time 1373.21, mean_step_time 1484.43, it/s 0.67
  1862. [0130 17:37:11 @multigpu.py:323] [p0115]  step: count(283), step_time 1393.71, mean_step_time 1375.52, it/s 0.73
  1863. [0130 17:37:11 @multigpu.py:323] [p0576]  step: count(306), step_time 1448.18, mean_step_time 1450.44, it/s 0.69
  1864. [0130 17:37:12 @multigpu.py:323] [p0574]  step: count(299), step_time 1393.04, mean_step_time 1483.69, it/s 0.67
  1865. [0130 17:37:12 @multigpu.py:323] [p0115]  step: count(284), step_time 1408.19, mean_step_time 1378.12, it/s 0.73
  1866. [0130 17:37:12 @multigpu.py:323] [p0576]  step: count(307), step_time 1367.6, mean_step_time 1448.74, it/s 0.69
  1867. [0130 17:37:13 @multigpu.py:323] [p0574]  step: count(300), step_time 1382.7, mean_step_time 1479.45, it/s 0.68
  1868. sending debugging info...
  1869. sending to address tcp://p0112:61216
  1870. ##### Sending to neptune:  mean_delay :  0.120217489137 , 0.0 #####
  1871. ##### Sending to neptune:  max_delay :  0.120217489137 , -0.0 #####
  1872. sending to address tcp://p0112:61216
  1873. ##### Sending to neptune:  min_delay :  0.120217489137 , -0.0 #####
  1874. [u'delays', [0.0, -0.0, -0.0]]
  1875. receiving
  1876. ##### Sending to neptune:  cost :  0.120218040546 , -0.0103861913085 #####
  1877. ##### Sending to neptune:  policy_loss :  0.120218040546 , -0.244617804885 #####
  1878. sending to address tcp://p0112:61216
  1879. ##### Sending to neptune:  xentropy_loss :  0.120218040546 , -2.28823947906 #####
  1880. ##### Sending to neptune:  value_loss :  0.120218040546 , 1.20342481136 #####
  1881. ##### Sending to neptune:  advantage :  0.120218040546 , 0.00103269040119 #####
  1882. ##### Sending to neptune:  pred_reward :  0.120218040546 , 0.395009964705 #####
  1883. ##### Sending to neptune:  max_logit :  0.120218040546 , 0.193438783288 #####
  1884. [u'loss', -0.010386191308498383, -0.24461780488491058, -2.2882394790649414, 1.2034248113632202, 0.0010326904011890292, 0.39500996470451355, 0.19343878328800201]
  1885. receiving
  1886. ##### Sending to neptune:  active_relus :  0.120218533013 , 8887224.41 #####
  1887. ##### Sending to neptune:  dp_per_s :  0.120218533013 , 89.5399337807 #####
  1888. [u'other', 8887224.41, 89.53993378070659]
  1889. receiving
  1890. [0130 17:37:13 @multigpu.py:323] [p0115]  step: count(285), step_time 1365.26, mean_step_time 1378.19, it/s 0.73
  1891. [0130 17:37:14 @multigpu.py:323] [p0576]  step: count(308), step_time 1443.34, mean_step_time 1447.1, it/s 0.69
  1892. [0130 17:37:15 @multigpu.py:323] [p0574]  step: count(301), step_time 1380.97, mean_step_time 1468.75, it/s 0.68
  1893. [0130 17:37:15 @multigpu.py:323] [p0115]  step: count(286), step_time 1412.84, mean_step_time 1378.96, it/s 0.73
  1894. [0130 17:37:15 @multigpu.py:323] [p0576]  step: count(309), step_time 1374.83, mean_step_time 1445.91, it/s 0.69
  1895. [0130 17:37:16 @multigpu.py:323] [p0574]  step: count(302), step_time 1371.75, mean_step_time 1458.32, it/s 0.69
  1896. [0130 17:37:16 @multigpu.py:323] [p0115]  step: count(287), step_time 1411.56, mean_step_time 1381.24, it/s 0.72
  1897. [0130 17:37:17 @multigpu.py:323] [p0576]  step: count(310), step_time 1376.81, mean_step_time 1448.46, it/s 0.69
  1898. [0130 17:37:18 @multigpu.py:323] [p0574]  step: count(303), step_time 1409.68, mean_step_time 1455.41, it/s 0.69
  1899. [0130 17:37:19 @multigpu.py:323] [p0574]  step: count(304), step_time 1364.59, mean_step_time 1442.68, it/s 0.69
  1900. [0130 17:37:19 @multigpu.py:323] [p0115]  step: count(288), step_time 2615.34, mean_step_time 1442.79, it/s 0.69
  1901. [0130 17:37:19 @multigpu.py:323] [p0576]  step: count(311), step_time 2319.64, mean_step_time 1442.63, it/s 0.69
  1902. [0130 17:37:20 @multigpu.py:323] [p0115]  step: count(289), step_time 1383.5, mean_step_time 1442.03, it/s 0.69
  1903. [0130 17:37:20 @multigpu.py:323] [p0576]  step: count(312), step_time 1439.56, mean_step_time 1446.74, it/s 0.69
  1904. [0130 17:37:20 @multigpu.py:323] [p0574]  step: count(305), step_time 1487.43, mean_step_time 1436.35, it/s 0.7
  1905. [0130 17:37:22 @multigpu.py:323] [p0115]  step: count(290), step_time 1381.82, mean_step_time 1443.98, it/s 0.69
  1906. [0130 17:37:22 @multigpu.py:323] [p0576]  step: count(313), step_time 1420.89, mean_step_time 1448.03, it/s 0.69
  1907. [0130 17:37:22 @multigpu.py:323] [p0574]  step: count(306), step_time 1430.76, mean_step_time 1426.36, it/s 0.7
  1908. [0130 17:37:23 @multigpu.py:323] [p0115]  step: count(291), step_time 1380.89, mean_step_time 1447.58, it/s 0.69
  1909. [0130 17:37:23 @multigpu.py:323] [p0576]  step: count(314), step_time 1397.24, mean_step_time 1450.58, it/s 0.69
  1910. [0130 17:37:23 @multigpu.py:323] [p0574]  step: count(307), step_time 1395.27, mean_step_time 1420.04, it/s 0.7
  1911. [0130 17:37:24 @multigpu.py:323] [p0115]  step: count(292), step_time 1402.06, mean_step_time 1447.06, it/s 0.69
  1912. [0130 17:37:25 @multigpu.py:323] [p0576]  step: count(315), step_time 1441.95, mean_step_time 1454.18, it/s 0.69
  1913. sending to address tcp://p0112:61216
  1914. ##### Sending to neptune:  online_score :  0.123340421054 , 1.1 #####
  1915. [u'online', 1.1]
  1916. ##### Sending to neptune:  active_workers :  0.123340500792 , 3 #####
  1917. receiving
  1918. [0130 17:37:25 @multigpu.py:323] [p0574]  step: count(308), step_time 1530.98, mean_step_time 1428.65, it/s 0.7
  1919. [0130 17:37:26 @multigpu.py:323] [p0115]  step: count(293), step_time 1416.82, mean_step_time 1448.6, it/s 0.69
  1920. [0130 17:37:26 @multigpu.py:323] [p0576]  step: count(316), step_time 1384.3, mean_step_time 1452.18, it/s 0.69
  1921. [0130 17:37:26 @multigpu.py:323] [p0574]  step: count(309), step_time 1488.91, mean_step_time 1433.66, it/s 0.7
  1922. [0130 17:37:27 @multigpu.py:323] [p0115]  step: count(294), step_time 1349.49, mean_step_time 1447.06, it/s 0.69
  1923. [0130 17:37:27 @multigpu.py:323] [p0576]  step: count(317), step_time 1357.59, mean_step_time 1451.62, it/s 0.69
  1924. [0130 17:37:28 @multigpu.py:323] [p0574]  step: count(310), step_time 1442.14, mean_step_time 1436.04, it/s 0.7
  1925. [0130 17:37:29 @multigpu.py:323] [p0115]  step: count(295), step_time 1392.08, mean_step_time 1447.68, it/s 0.69
  1926. [0130 17:37:29 @multigpu.py:323] [p0576]  step: count(318), step_time 1462.51, mean_step_time 1455.0, it/s 0.69
  1927. [0130 17:37:29 @multigpu.py:323] [p0574]  step: count(311), step_time 1438.97, mean_step_time 1438.62, it/s 0.7
  1928. [0130 17:37:30 @multigpu.py:323] [p0115]  step: count(296), step_time 1361.63, mean_step_time 1446.41, it/s 0.69
  1929. [0130 17:37:30 @multigpu.py:323] [p0576]  step: count(319), step_time 1387.51, mean_step_time 1455.09, it/s 0.69
  1930. [0130 17:37:31 @multigpu.py:323] [p0574]  step: count(312), step_time 1377.31, mean_step_time 1438.99, it/s 0.69
  1931. [0130 17:37:31 @multigpu.py:323] [p0115]  step: count(297), step_time 1405.26, mean_step_time 1447.47, it/s 0.69
  1932. [0130 17:37:32 @multigpu.py:323] [p0576]  step: count(320), step_time 1379.68, mean_step_time 1450.96, it/s 0.69
  1933. [0130 17:37:32 @multigpu.py:323] [p0574]  step: count(313), step_time 1417.67, mean_step_time 1440.93, it/s 0.69
  1934. [0130 17:37:33 @multigpu.py:323] [p0115]  step: count(298), step_time 1410.03, mean_step_time 1449.9, it/s 0.69
  1935. [0130 17:37:33 @multigpu.py:323] [p0576]  step: count(321), step_time 1376.52, mean_step_time 1448.66, it/s 0.69
  1936. [0130 17:37:33 @multigpu.py:323] [p0574]  step: count(314), step_time 1339.04, mean_step_time 1431.68, it/s 0.7
  1937. [0130 17:37:34 @multigpu.py:323] [p0115]  step: count(299), step_time 1389.03, mean_step_time 1449.82, it/s 0.69
  1938. [0130 17:37:34 @multigpu.py:323] [p0576]  step: count(322), step_time 1386.44, mean_step_time 1449.29, it/s 0.69
  1939. [0130 17:37:35 @multigpu.py:323] [p0574]  step: count(315), step_time 1398.39, mean_step_time 1431.79, it/s 0.7
  1940. [0130 17:37:36 @multigpu.py:323] [p0115]  step: count(300), step_time 1443.03, mean_step_time 1453.65, it/s 0.69
  1941. sending debugging info...
  1942. sending to address tcp://p0112:61216
  1943. ##### Sending to neptune:  mean_delay :  0.126398886376 , 0.0 #####
  1944. sending to address tcp://p0112:61216
  1945. ##### Sending to neptune:  max_delay :  0.126398886376 , -0.0 #####
  1946. ##### Sending to neptune:  min_delay :  0.126398886376 , -0.0 #####
  1947. [u'delays', [0.0, -0.0, -0.0]]
  1948. receiving
  1949. ##### Sending to neptune:  cost :  0.126399410831 , -0.0113957431167 #####
  1950. sending to address tcp://p0112:61216
  1951. ##### Sending to neptune:  policy_loss :  0.126399410831 , -0.226604878902 #####
  1952. ##### Sending to neptune:  xentropy_loss :  0.126399410831 , -2.28844261169 #####
  1953. ##### Sending to neptune:  value_loss :  0.126399410831 , 1.05639243126 #####
  1954. ##### Sending to neptune:  advantage :  0.126399410831 , 0.00103650894016 #####
  1955. ##### Sending to neptune:  pred_reward :  0.126399410831 , 0.386131823063 #####
  1956. ##### Sending to neptune:  max_logit :  0.126399410831 , 0.19203697145 #####
  1957. [u'loss', -0.011395743116736412, -0.2266048789024353, -2.288442611694336, 1.0563924312591553, 0.0010365089401602745, 0.38613182306289673, 0.192036971449852]
  1958. receiving
  1959. ##### Sending to neptune:  active_relus :  0.126399871376 , 8856618.33 #####
  1960. ##### Sending to neptune:  dp_per_s :  0.126399871376 , 91.271358836 #####
  1961. [u'other', 8856618.33, 91.27135883601044]
  1962. receiving
  1963. [0130 17:37:36 @multigpu.py:323] [p0576]  step: count(323), step_time 1405.49, mean_step_time 1447.36, it/s 0.69
  1964. [0130 17:37:36 @multigpu.py:323] [p0574]  step: count(316), step_time 1407.92, mean_step_time 1420.67, it/s 0.7
  1965. [0130 17:37:37 @multigpu.py:323] [p0115]  step: count(301), step_time 1387.53, mean_step_time 1454.42, it/s 0.69
  1966. [0130 17:37:37 @multigpu.py:323] [p0576]  step: count(324), step_time 1378.84, mean_step_time 1448.78, it/s 0.69
  1967. [0130 17:37:37 @multigpu.py:323] [p0574]  step: count(317), step_time 1393.47, mean_step_time 1411.21, it/s 0.71
  1968. sending to address tcp://p0112:61216
  1969. ##### Sending to neptune:  online_score :  0.12702402691 , 0.9 #####
  1970. [u'online', 0.9]
  1971. receiving
  1972. sending to address tcp://p0112:61216
  1973. ##### Sending to neptune:  online_score :  0.127058637738 , 1.1 #####
  1974. [u'online', 1.1]
  1975. receiving
  1976. [0130 17:37:38 @multigpu.py:323] [p0115]  step: count(302), step_time 1390.81, mean_step_time 1455.04, it/s 0.69
  1977. [0130 17:37:39 @multigpu.py:323] [p0576]  step: count(325), step_time 1451.99, mean_step_time 1450.05, it/s 0.69
  1978. [0130 17:37:39 @multigpu.py:323] [p0574]  step: count(318), step_time 1425.59, mean_step_time 1413.83, it/s 0.71
  1979. [0130 17:37:40 @multigpu.py:323] [p0115]  step: count(303), step_time 1408.57, mean_step_time 1455.79, it/s 0.69
  1980. [0130 17:37:40 @multigpu.py:323] [p0576]  step: count(326), step_time 1362.75, mean_step_time 1445.77, it/s 0.69
  1981. [0130 17:37:40 @multigpu.py:323] [p0574]  step: count(319), step_time 1372.05, mean_step_time 1412.78, it/s 0.71
  1982. [0130 17:37:41 @multigpu.py:323] [p0115]  step: count(304), step_time 1345.33, mean_step_time 1452.64, it/s 0.69
  1983. [0130 17:37:41 @multigpu.py:323] [p0576]  step: count(327), step_time 1445.58, mean_step_time 1449.67, it/s 0.69
  1984. [0130 17:37:42 @multigpu.py:323] [p0574]  step: count(320), step_time 1361.26, mean_step_time 1411.71, it/s 0.71
  1985. [0130 17:37:43 @multigpu.py:323] [p0115]  step: count(305), step_time 1437.42, mean_step_time 1456.25, it/s 0.69
  1986. [0130 17:37:43 @multigpu.py:323] [p0576]  step: count(328), step_time 1383.55, mean_step_time 1446.68, it/s 0.69
  1987. [0130 17:37:43 @multigpu.py:323] [p0574]  step: count(321), step_time 1381.44, mean_step_time 1411.73, it/s 0.71
  1988. [0130 17:37:44 @multigpu.py:323] [p0115]  step: count(306), step_time 1359.95, mean_step_time 1453.61, it/s 0.69
  1989. [0130 17:37:44 @multigpu.py:323] [p0576]  step: count(329), step_time 1418.87, mean_step_time 1448.89, it/s 0.69
  1990. [0130 17:37:44 @multigpu.py:323] [p0574]  step: count(322), step_time 1415.14, mean_step_time 1413.9, it/s 0.71
  1991. [0130 17:37:45 @multigpu.py:323] [p0115]  step: count(307), step_time 1396.46, mean_step_time 1452.85, it/s 0.69
  1992. [0130 17:37:46 @multigpu.py:323] [p0576]  step: count(330), step_time 1455.6, mean_step_time 1452.83, it/s 0.69
  1993. [0130 17:37:46 @multigpu.py:323] [p0574]  step: count(323), step_time 1414.92, mean_step_time 1414.16, it/s 0.71
  1994. [0130 17:37:47 @multigpu.py:323] [p0574]  step: count(324), step_time 1423.54, mean_step_time 1417.11, it/s 0.71
  1995. [0130 17:37:47 @multigpu.py:323] [p0115]  step: count(308), step_time 1900.99, mean_step_time 1417.14, it/s 0.71
  1996. [0130 17:37:47 @multigpu.py:323] [p0576]  step: count(331), step_time 1604.34, mean_step_time 1417.06, it/s 0.71
  1997. [0130 17:37:49 @multigpu.py:323] [p0115]  step: count(309), step_time 1365.43, mean_step_time 1416.23, it/s 0.71
  1998. [0130 17:37:49 @multigpu.py:323] [p0574]  step: count(325), step_time 1368.14, mean_step_time 1411.15, it/s 0.71
  1999. [0130 17:37:49 @multigpu.py:323] [p0576]  step: count(332), step_time 1407.4, mean_step_time 1415.45, it/s 0.71
  2000. [0130 17:37:50 @multigpu.py:323] [p0574]  step: count(326), step_time 1396.46, mean_step_time 1409.43, it/s 0.71
  2001. [0130 17:37:50 @multigpu.py:323] [p0115]  step: count(310), step_time 1424.3, mean_step_time 1418.36, it/s 0.71
  2002. [0130 17:37:50 @multigpu.py:323] [p0576]  step: count(333), step_time 1397.06, mean_step_time 1414.26, it/s 0.71
  2003. [0130 17:37:51 @multigpu.py:323] [p0574]  step: count(327), step_time 1358.35, mean_step_time 1407.58, it/s 0.71
  2004. [0130 17:37:51 @multigpu.py:323] [p0115]  step: count(311), step_time 1356.57, mean_step_time 1417.14, it/s 0.71
  2005. [0130 17:37:51 @multigpu.py:323] [p0576]  step: count(334), step_time 1354.64, mean_step_time 1412.13, it/s 0.71
  2006. [0130 17:37:53 @multigpu.py:323] [p0574]  step: count(328), step_time 1380.41, mean_step_time 1400.06, it/s 0.71
  2007. [0130 17:37:53 @multigpu.py:323] [p0115]  step: count(312), step_time 1374.11, mean_step_time 1415.74, it/s 0.71
  2008. [0130 17:37:53 @multigpu.py:323] [p0576]  step: count(335), step_time 1438.54, mean_step_time 1411.96, it/s 0.71
  2009. [0130 17:37:54 @multigpu.py:323] [p0574]  step: count(329), step_time 1412.16, mean_step_time 1396.22, it/s 0.72
  2010. [0130 17:37:54 @multigpu.py:323] [p0115]  step: count(313), step_time 1416.59, mean_step_time 1415.73, it/s 0.71
  2011. [0130 17:37:54 @multigpu.py:323] [p0576]  step: count(336), step_time 1447.87, mean_step_time 1415.14, it/s 0.71
  2012. sending to address tcp://p0112:61216
  2013. ##### Sending to neptune:  online_score :  0.131745723883 , 2.1 #####
  2014. [u'online', 2.1]
  2015. receiving
  2016. [0130 17:37:56 @multigpu.py:323] [p0574]  step: count(330), step_time 1360.8, mean_step_time 1392.15, it/s 0.72
  2017. [0130 17:37:56 @multigpu.py:323] [p0115]  step: count(314), step_time 1370.28, mean_step_time 1416.77, it/s 0.71
  2018. [0130 17:37:56 @multigpu.py:323] [p0576]  step: count(337), step_time 1417.26, mean_step_time 1418.12, it/s 0.71
  2019. [0130 17:37:57 @multigpu.py:323] [p0574]  step: count(331), step_time 1391.51, mean_step_time 1389.78, it/s 0.72
  2020. [0130 17:37:57 @multigpu.py:323] [p0115]  step: count(315), step_time 1387.7, mean_step_time 1416.55, it/s 0.71
  2021. [0130 17:37:57 @multigpu.py:323] [p0576]  step: count(338), step_time 1444.41, mean_step_time 1417.22, it/s 0.71
  2022. [0130 17:37:58 @multigpu.py:323] [p0574]  step: count(332), step_time 1374.04, mean_step_time 1389.62, it/s 0.72
  2023. [0130 17:37:58 @multigpu.py:323] [p0115]  step: count(316), step_time 1388.04, mean_step_time 1417.87, it/s 0.71
  2024. [0130 17:37:59 @multigpu.py:323] [p0576]  step: count(339), step_time 1442.65, mean_step_time 1419.97, it/s 0.7
  2025. sending to address tcp://p0112:61216
  2026. ##### Sending to neptune:  online_score :  0.132890012993 , 1.4 #####
  2027. [u'online', 1.4]
  2028. receiving
  2029. [0130 17:38:00 @multigpu.py:323] [p0574]  step: count(333), step_time 1416.93, mean_step_time 1389.58, it/s 0.72
  2030. [0130 17:38:00 @multigpu.py:323] [p0115]  step: count(317), step_time 1383.89, mean_step_time 1416.8, it/s 0.71
  2031. [0130 17:38:00 @multigpu.py:323] [p0576]  step: count(340), step_time 1449.19, mean_step_time 1423.45, it/s 0.7
  2032. [0130 17:38:01 @multigpu.py:323] [p0574]  step: count(334), step_time 1360.92, mean_step_time 1390.67, it/s 0.72
  2033. [0130 17:38:01 @multigpu.py:323] [p0115]  step: count(318), step_time 1384.03, mean_step_time 1415.5, it/s 0.71
  2034. [0130 17:38:01 @multigpu.py:323] [p0576]  step: count(341), step_time 1400.62, mean_step_time 1424.65, it/s 0.7
  2035. sending to address tcp://p0112:61216
  2036. ##### Sending to neptune:  online_score :  0.133605820537 , 1.9 #####
  2037. [u'online', 1.9]
  2038. receiving
  2039. [0130 17:38:02 @multigpu.py:323] [p0115]  step: count(319), step_time 1350.75, mean_step_time 1413.59, it/s 0.71
  2040. [0130 17:38:03 @multigpu.py:323] [p0574]  step: count(335), step_time 1407.67, mean_step_time 1391.14, it/s 0.72
  2041. [0130 17:38:03 @multigpu.py:323] [p0576]  step: count(342), step_time 1447.8, mean_step_time 1427.72, it/s 0.7
  2042. [0130 17:38:04 @multigpu.py:323] [p0574]  step: count(336), step_time 1384.51, mean_step_time 1389.97, it/s 0.72
  2043. [0130 17:38:04 @multigpu.py:323] [p0115]  step: count(320), step_time 1447.59, mean_step_time 1413.82, it/s 0.71
  2044. [0130 17:38:04 @multigpu.py:323] [p0576]  step: count(343), step_time 1418.6, mean_step_time 1428.38, it/s 0.7
  2045. [0130 17:38:05 @multigpu.py:323] [p0574]  step: count(337), step_time 1368.34, mean_step_time 1388.71, it/s 0.72
  2046. [0130 17:38:05 @multigpu.py:323] [p0115]  step: count(321), step_time 1407.05, mean_step_time 1414.79, it/s 0.71
  2047. [0130 17:38:06 @multigpu.py:323] [p0576]  step: count(344), step_time 1410.33, mean_step_time 1429.95, it/s 0.7
  2048. [0130 17:38:07 @multigpu.py:323] [p0574]  step: count(338), step_time 1393.57, mean_step_time 1387.11, it/s 0.72
  2049. [0130 17:38:07 @multigpu.py:323] [p0115]  step: count(322), step_time 1391.66, mean_step_time 1414.83, it/s 0.71
  2050. [0130 17:38:07 @multigpu.py:323] [p0576]  step: count(345), step_time 1421.3, mean_step_time 1428.42, it/s 0.7
  2051. [0130 17:38:08 @multigpu.py:323] [p0574]  step: count(339), step_time 1349.97, mean_step_time 1386.0, it/s 0.72
  2052. [0130 17:38:08 @multigpu.py:323] [p0115]  step: count(323), step_time 1379.38, mean_step_time 1413.38, it/s 0.71
  2053. [0130 17:38:09 @multigpu.py:323] [p0576]  step: count(346), step_time 1406.67, mean_step_time 1430.61, it/s 0.7
  2054. [0130 17:38:09 @multigpu.py:323] [p0574]  step: count(340), step_time 1452.74, mean_step_time 1390.58, it/s 0.72
  2055. [0130 17:38:09 @multigpu.py:323] [p0115]  step: count(324), step_time 1370.54, mean_step_time 1414.64, it/s 0.71
  2056. [0130 17:38:10 @multigpu.py:323] [p0576]  step: count(347), step_time 1397.1, mean_step_time 1428.19, it/s 0.7
  2057. [0130 17:38:11 @multigpu.py:323] [p0115]  step: count(325), step_time 1345.95, mean_step_time 1410.06, it/s 0.71
  2058. [0130 17:38:11 @multigpu.py:323] [p0574]  step: count(341), step_time 1485.56, mean_step_time 1395.78, it/s 0.72
  2059. [0130 17:38:11 @multigpu.py:323] [p0576]  step: count(348), step_time 1434.95, mean_step_time 1430.76, it/s 0.7
  2060. [0130 17:38:12 @multigpu.py:323] [p0115]  step: count(326), step_time 1372.42, mean_step_time 1410.69, it/s 0.71
  2061. [0130 17:38:12 @multigpu.py:323] [p0574]  step: count(342), step_time 1347.73, mean_step_time 1392.41, it/s 0.72
  2062. [0130 17:38:13 @multigpu.py:323] [p0576]  step: count(349), step_time 1477.29, mean_step_time 1433.68, it/s 0.7
  2063. [0130 17:38:14 @multigpu.py:323] [p0115]  step: count(327), step_time 1419.62, mean_step_time 1411.84, it/s 0.71
  2064. [0130 17:38:14 @multigpu.py:323] [p0574]  step: count(343), step_time 1402.41, mean_step_time 1391.79, it/s 0.72
  2065. [0130 17:38:14 @multigpu.py:323] [p0576]  step: count(350), step_time 1375.61, mean_step_time 1429.68, it/s 0.7
  2066. [0130 17:38:16 @multigpu.py:323] [p0576]  step: count(351), step_time 1423.33, mean_step_time 1420.63, it/s 0.7
  2067. [0130 17:38:16 @multigpu.py:323] [p0574]  step: count(344), step_time 2003.9, mean_step_time 1420.81, it/s 0.7
  2068. [0130 17:38:16 @multigpu.py:323] [p0115]  step: count(328), step_time 2080.96, mean_step_time 1420.84, it/s 0.7
  2069. [0130 17:38:17 @multigpu.py:323] [p0115]  step: count(329), step_time 1369.82, mean_step_time 1421.06, it/s 0.7
  2070. [0130 17:38:17 @multigpu.py:323] [p0574]  step: count(345), step_time 1407.39, mean_step_time 1422.77, it/s 0.7
  2071. [0130 17:38:17 @multigpu.py:323] [p0576]  step: count(352), step_time 1426.08, mean_step_time 1421.56, it/s 0.7
  2072. [0130 17:38:18 @multigpu.py:323] [p0574]  step: count(346), step_time 1367.47, mean_step_time 1421.32, it/s 0.7
  2073. [0130 17:38:18 @multigpu.py:323] [p0115]  step: count(330), step_time 1413.16, mean_step_time 1420.5, it/s 0.7
  2074. [0130 17:38:19 @multigpu.py:323] [p0576]  step: count(353), step_time 1400.77, mean_step_time 1421.75, it/s 0.7
  2075. sending to address tcp://p0112:61216
  2076. ##### Sending to neptune:  online_score :  0.138402113583 , 0.7 #####
  2077. [u'online', 0.7]
  2078. receiving
  2079. [0130 17:38:20 @multigpu.py:323] [p0574]  step: count(347), step_time 1365.51, mean_step_time 1421.68, it/s 0.7
  2080. [0130 17:38:20 @multigpu.py:323] [p0115]  step: count(331), step_time 1395.58, mean_step_time 1422.46, it/s 0.7
  2081. [0130 17:38:20 @multigpu.py:323] [p0576]  step: count(354), step_time 1377.34, mean_step_time 1422.89, it/s 0.7
  2082. [0130 17:38:21 @multigpu.py:323] [p0574]  step: count(348), step_time 1433.04, mean_step_time 1424.31, it/s 0.7
  2083. [0130 17:38:21 @multigpu.py:323] [p0115]  step: count(332), step_time 1404.26, mean_step_time 1423.96, it/s 0.7
  2084. [0130 17:38:21 @multigpu.py:323] [p0576]  step: count(355), step_time 1390.11, mean_step_time 1420.46, it/s 0.7
  2085. [0130 17:38:23 @multigpu.py:323] [p0115]  step: count(333), step_time 1366.41, mean_step_time 1421.45, it/s 0.7
  2086. [0130 17:38:23 @multigpu.py:323] [p0574]  step: count(349), step_time 1398.66, mean_step_time 1423.63, it/s 0.7
  2087. [0130 17:38:23 @multigpu.py:323] [p0576]  step: count(356), step_time 1410.44, mean_step_time 1418.59, it/s 0.7
  2088. [0130 17:38:24 @multigpu.py:323] [p0574]  step: count(350), step_time 1343.51, mean_step_time 1422.77, it/s 0.7
  2089. [0130 17:38:24 @multigpu.py:323] [p0115]  step: count(334), step_time 1460.58, mean_step_time 1425.97, it/s 0.7
  2090. [0130 17:38:24 @multigpu.py:323] [p0576]  step: count(357), step_time 1432.73, mean_step_time 1419.37, it/s 0.7
  2091. sending to address tcp://p0112:61216
  2092. ##### Sending to neptune:  online_score :  0.139930734701 , 0.2 #####
  2093. [u'online', 0.2]
  2094. receiving
  2095. [0130 17:38:25 @multigpu.py:323] [p0574]  step: count(351), step_time 1363.59, mean_step_time 1421.37, it/s 0.7
  2096. [0130 17:38:25 @multigpu.py:323] [p0115]  step: count(335), step_time 1355.52, mean_step_time 1424.36, it/s 0.7
  2097. [0130 17:38:26 @multigpu.py:323] [p0576]  step: count(358), step_time 1470.23, mean_step_time 1420.66, it/s 0.7
  2098. [0130 17:38:27 @multigpu.py:323] [p0574]  step: count(352), step_time 1403.23, mean_step_time 1422.83, it/s 0.7
  2099. [0130 17:38:27 @multigpu.py:323] [p0115]  step: count(336), step_time 1392.58, mean_step_time 1424.59, it/s 0.7
  2100. [0130 17:38:27 @multigpu.py:323] [p0576]  step: count(359), step_time 1365.14, mean_step_time 1416.78, it/s 0.71
  2101. sending to address tcp://p0112:61216
  2102. ##### Sending to neptune:  online_score :  0.140695282751 , 1.5 #####
  2103. [u'online', 1.5]
  2104. ##### Sending to neptune:  active_workers :  0.140695316063 , 3 #####
  2105. receiving
  2106. [0130 17:38:28 @multigpu.py:323] [p0574]  step: count(353), step_time 1368.95, mean_step_time 1420.43, it/s 0.7
  2107. [0130 17:38:28 @multigpu.py:323] [p0115]  step: count(337), step_time 1372.64, mean_step_time 1424.03, it/s 0.7
  2108. [0130 17:38:28 @multigpu.py:323] [p0576]  step: count(360), step_time 1449.26, mean_step_time 1416.78, it/s 0.71
  2109. [0130 17:38:30 @multigpu.py:323] [p0574]  step: count(354), step_time 1349.47, mean_step_time 1419.86, it/s 0.7
  2110. [0130 17:38:30 @multigpu.py:323] [p0115]  step: count(338), step_time 1391.57, mean_step_time 1424.4, it/s 0.7
  2111. [0130 17:38:30 @multigpu.py:323] [p0576]  step: count(361), step_time 1403.23, mean_step_time 1416.92, it/s 0.71
  2112. [0130 17:38:31 @multigpu.py:323] [p0574]  step: count(355), step_time 1500.65, mean_step_time 1424.51, it/s 0.7
  2113. [0130 17:38:31 @multigpu.py:323] [p0115]  step: count(339), step_time 1386.29, mean_step_time 1426.18, it/s 0.7
  2114. [0130 17:38:31 @multigpu.py:323] [p0576]  step: count(362), step_time 1423.1, mean_step_time 1415.68, it/s 0.71
  2115. [0130 17:38:32 @multigpu.py:323] [p0574]  step: count(356), step_time 1366.82, mean_step_time 1423.63, it/s 0.7
  2116. [0130 17:38:32 @multigpu.py:323] [p0115]  step: count(340), step_time 1403.96, mean_step_time 1424.0, it/s 0.7
  2117. [0130 17:38:33 @multigpu.py:323] [p0576]  step: count(363), step_time 1424.65, mean_step_time 1415.98, it/s 0.71
  2118. [0130 17:38:34 @multigpu.py:323] [p0115]  step: count(341), step_time 1344.14, mean_step_time 1420.85, it/s 0.7
  2119. [0130 17:38:34 @multigpu.py:323] [p0574]  step: count(357), step_time 1401.42, mean_step_time 1425.28, it/s 0.7
  2120. [0130 17:38:34 @multigpu.py:323] [p0576]  step: count(364), step_time 1440.01, mean_step_time 1417.47, it/s 0.71
  2121. [0130 17:38:35 @multigpu.py:323] [p0115]  step: count(342), step_time 1358.81, mean_step_time 1419.21, it/s 0.7
  2122. [0130 17:38:35 @multigpu.py:323] [p0574]  step: count(358), step_time 1376.03, mean_step_time 1424.4, it/s 0.7
  2123. [0130 17:38:36 @multigpu.py:323] [p0576]  step: count(365), step_time 1424.0, mean_step_time 1417.6, it/s 0.71
  2124. [0130 17:38:37 @multigpu.py:323] [p0574]  step: count(359), step_time 1362.57, mean_step_time 1425.03, it/s 0.7
  2125. [0130 17:38:37 @multigpu.py:323] [p0115]  step: count(343), step_time 1408.86, mean_step_time 1420.68, it/s 0.7
  2126. [0130 17:38:37 @multigpu.py:323] [p0576]  step: count(366), step_time 1437.07, mean_step_time 1419.12, it/s 0.7
  2127. [0130 17:38:38 @multigpu.py:323] [p0574]  step: count(360), step_time 1373.87, mean_step_time 1421.09, it/s 0.7
  2128. [0130 17:38:38 @multigpu.py:323] [p0115]  step: count(344), step_time 1415.34, mean_step_time 1422.92, it/s 0.7
  2129. [0130 17:38:38 @multigpu.py:323] [p0576]  step: count(367), step_time 1466.44, mean_step_time 1422.59, it/s 0.7
  2130. [0130 17:38:39 @multigpu.py:323] [p0574]  step: count(361), step_time 1423.31, mean_step_time 1417.98, it/s 0.71
  2131. [0130 17:38:39 @multigpu.py:323] [p0115]  step: count(345), step_time 1372.49, mean_step_time 1424.25, it/s 0.7
  2132. [0130 17:38:40 @multigpu.py:323] [p0576]  step: count(368), step_time 1395.91, mean_step_time 1420.64, it/s 0.7
  2133. [0130 17:38:41 @multigpu.py:323] [p0115]  step: count(346), step_time 1396.56, mean_step_time 1425.46, it/s 0.7
  2134. [0130 17:38:41 @multigpu.py:323] [p0574]  step: count(362), step_time 1432.84, mean_step_time 1422.23, it/s 0.7
  2135. [0130 17:38:41 @multigpu.py:323] [p0576]  step: count(369), step_time 1425.34, mean_step_time 1418.04, it/s 0.71
  2136. [0130 17:38:42 @multigpu.py:323] [p0574]  step: count(363), step_time 1337.44, mean_step_time 1418.98, it/s 0.7
  2137. [0130 17:38:42 @multigpu.py:323] [p0115]  step: count(347), step_time 1367.95, mean_step_time 1422.87, it/s 0.7
  2138. [0130 17:38:43 @multigpu.py:323] [p0576]  step: count(370), step_time 1406.6, mean_step_time 1419.59, it/s 0.7
  2139. [0130 17:38:44 @multigpu.py:323] [p0576]  step: count(371), step_time 1513.52, mean_step_time 1424.1, it/s 0.7
  2140. [0130 17:38:44 @multigpu.py:323] [p0115]  step: count(348), step_time 2106.69, mean_step_time 1424.16, it/s 0.7
  2141. [0130 17:38:44 @multigpu.py:323] [p0574]  step: count(364), step_time 2107.26, mean_step_time 1424.15, it/s 0.7
  2142. sending to address tcp://p0112:61216
  2143. ##### Sending to neptune:  online_score :  0.145541653037 , 2.1 #####
  2144. [u'online', 2.1]
  2145. receiving
  2146. [0130 17:38:46 @multigpu.py:323] [p0574]  step: count(365), step_time 1374.01, mean_step_time 1422.48, it/s 0.7
  2147. [0130 17:38:46 @multigpu.py:323] [p0576]  step: count(372), step_time 1403.78, mean_step_time 1422.98, it/s 0.7
  2148. [0130 17:38:46 @multigpu.py:323] [p0115]  step: count(349), step_time 1428.43, mean_step_time 1427.09, it/s 0.7
  2149. [0130 17:38:47 @multigpu.py:323] [p0115]  step: count(350), step_time 1381.07, mean_step_time 1425.49, it/s 0.7
  2150. [0130 17:38:47 @multigpu.py:323] [p0574]  step: count(366), step_time 1463.64, mean_step_time 1427.29, it/s 0.7
  2151. [0130 17:38:47 @multigpu.py:323] [p0576]  step: count(373), step_time 1461.16, mean_step_time 1426.0, it/s 0.7
  2152. [0130 17:38:48 @multigpu.py:323] [p0115]  step: count(351), step_time 1376.54, mean_step_time 1424.53, it/s 0.7
  2153. [0130 17:38:48 @multigpu.py:323] [p0574]  step: count(367), step_time 1424.99, mean_step_time 1430.27, it/s 0.7
  2154. [0130 17:38:48 @multigpu.py:323] [p0576]  step: count(374), step_time 1417.18, mean_step_time 1427.99, it/s 0.7
  2155. sending to address tcp://p0112:61216
  2156. ##### Sending to neptune:  online_score :  0.146764651073 , 1.8 #####
  2157. [u'online', 1.8]
  2158. receiving
  2159. sending to address tcp://p0112:61216
  2160. ##### Sending to neptune:  online_score :  0.146777135266 , 0.9 #####
  2161. [u'online', 0.9]
  2162. receiving
  2163. [0130 17:38:50 @multigpu.py:323] [p0115]  step: count(352), step_time 1377.77, mean_step_time 1423.21, it/s 0.7
  2164. [0130 17:38:50 @multigpu.py:323] [p0574]  step: count(368), step_time 1360.8, mean_step_time 1426.65, it/s 0.7
  2165. [0130 17:38:50 @multigpu.py:323] [p0576]  step: count(375), step_time 1404.25, mean_step_time 1428.7, it/s 0.7
  2166. [0130 17:38:51 @multigpu.py:323] [p0115]  step: count(353), step_time 1405.07, mean_step_time 1425.14, it/s 0.7
  2167. [0130 17:38:51 @multigpu.py:323] [p0574]  step: count(369), step_time 1383.47, mean_step_time 1425.89, it/s 0.7
  2168. [0130 17:38:51 @multigpu.py:323] [p0576]  step: count(376), step_time 1468.44, mean_step_time 1431.6, it/s 0.7
  2169. [0130 17:38:53 @multigpu.py:323] [p0115]  step: count(354), step_time 1378.63, mean_step_time 1421.05, it/s 0.7
  2170. [0130 17:38:53 @multigpu.py:323] [p0574]  step: count(370), step_time 1375.83, mean_step_time 1427.51, it/s 0.7
  2171. [0130 17:38:53 @multigpu.py:323] [p0576]  step: count(377), step_time 1402.5, mean_step_time 1430.09, it/s 0.7
  2172. [0130 17:38:54 @multigpu.py:323] [p0574]  step: count(371), step_time 1368.96, mean_step_time 1427.78, it/s 0.7
  2173. [0130 17:38:54 @multigpu.py:323] [p0115]  step: count(355), step_time 1424.46, mean_step_time 1424.49, it/s 0.7
  2174. [0130 17:38:54 @multigpu.py:323] [p0576]  step: count(378), step_time 1438.33, mean_step_time 1428.5, it/s 0.7
  2175. [0130 17:38:55 @multigpu.py:323] [p0115]  step: count(356), step_time 1400.95, mean_step_time 1424.91, it/s 0.7
  2176. [0130 17:38:55 @multigpu.py:323] [p0574]  step: count(372), step_time 1506.86, mean_step_time 1432.96, it/s 0.7
  2177. [0130 17:38:56 @multigpu.py:323] [p0576]  step: count(379), step_time 1412.34, mean_step_time 1430.86, it/s 0.7
  2178. [0130 17:38:57 @multigpu.py:323] [p0115]  step: count(357), step_time 1399.37, mean_step_time 1426.25, it/s 0.7
  2179. [0130 17:38:57 @multigpu.py:323] [p0574]  step: count(373), step_time 1384.96, mean_step_time 1433.76, it/s 0.7
  2180. [0130 17:38:57 @multigpu.py:323] [p0576]  step: count(380), step_time 1378.29, mean_step_time 1427.31, it/s 0.7
  2181. [0130 17:38:58 @multigpu.py:323] [p0115]  step: count(358), step_time 1350.07, mean_step_time 1424.17, it/s 0.7
  2182. [0130 17:38:58 @multigpu.py:323] [p0574]  step: count(374), step_time 1464.46, mean_step_time 1439.51, it/s 0.69
  2183. [0130 17:38:58 @multigpu.py:323] [p0576]  step: count(381), step_time 1442.18, mean_step_time 1429.25, it/s 0.7
  2184. [0130 17:39:00 @multigpu.py:323] [p0115]  step: count(359), step_time 1409.5, mean_step_time 1425.33, it/s 0.7
  2185. [0130 17:39:00 @multigpu.py:323] [p0574]  step: count(375), step_time 1375.47, mean_step_time 1433.25, it/s 0.7
  2186. [0130 17:39:00 @multigpu.py:323] [p0576]  step: count(382), step_time 1397.03, mean_step_time 1427.95, it/s 0.7
  2187. [0130 17:39:01 @multigpu.py:323] [p0115]  step: count(360), step_time 1412.0, mean_step_time 1425.74, it/s 0.7
  2188. [0130 17:39:01 @multigpu.py:323] [p0574]  step: count(376), step_time 1413.74, mean_step_time 1435.6, it/s 0.7
  2189. [0130 17:39:01 @multigpu.py:323] [p0576]  step: count(383), step_time 1439.07, mean_step_time 1428.67, it/s 0.7
  2190. [0130 17:39:02 @multigpu.py:323] [p0115]  step: count(361), step_time 1429.66, mean_step_time 1430.01, it/s 0.7
  2191. [0130 17:39:02 @multigpu.py:323] [p0574]  step: count(377), step_time 1388.78, mean_step_time 1434.96, it/s 0.7
  2192. [0130 17:39:03 @multigpu.py:323] [p0576]  step: count(384), step_time 1395.49, mean_step_time 1426.45, it/s 0.7
  2193. [0130 17:39:04 @multigpu.py:323] [p0115]  step: count(362), step_time 1366.82, mean_step_time 1430.41, it/s 0.7
  2194. [0130 17:39:04 @multigpu.py:323] [p0574]  step: count(378), step_time 1380.51, mean_step_time 1435.19, it/s 0.7
  2195. [0130 17:39:04 @multigpu.py:323] [p0576]  step: count(385), step_time 1405.33, mean_step_time 1425.51, it/s 0.7
  2196. [0130 17:39:05 @multigpu.py:323] [p0115]  step: count(363), step_time 1430.37, mean_step_time 1431.49, it/s 0.7
  2197. [0130 17:39:05 @multigpu.py:323] [p0574]  step: count(379), step_time 1398.73, mean_step_time 1437.0, it/s 0.7
  2198. [0130 17:39:06 @multigpu.py:323] [p0576]  step: count(386), step_time 1445.15, mean_step_time 1425.92, it/s 0.7
  2199. sending to address tcp://p0112:61216
  2200. ##### Sending to neptune:  online_score :  0.151504749722 , 0.8 #####
  2201. [u'online', 0.8]
  2202. receiving
  2203. [0130 17:39:07 @multigpu.py:323] [p0115]  step: count(364), step_time 1380.69, mean_step_time 1429.75, it/s 0.7
  2204. [0130 17:39:07 @multigpu.py:323] [p0574]  step: count(380), step_time 1397.96, mean_step_time 1438.2, it/s 0.7
  2205. [0130 17:39:07 @multigpu.py:323] [p0576]  step: count(387), step_time 1396.97, mean_step_time 1422.44, it/s 0.7
  2206. [0130 17:39:08 @multigpu.py:323] [p0115]  step: count(365), step_time 1369.8, mean_step_time 1429.62, it/s 0.7
  2207. [0130 17:39:08 @multigpu.py:323] [p0574]  step: count(381), step_time 1393.46, mean_step_time 1436.71, it/s 0.7
  2208. [0130 17:39:08 @multigpu.py:323] [p0576]  step: count(388), step_time 1398.0, mean_step_time 1422.55, it/s 0.7
  2209. [0130 17:39:09 @multigpu.py:323] [p0115]  step: count(366), step_time 1370.75, mean_step_time 1428.33, it/s 0.7
  2210. [0130 17:39:09 @multigpu.py:323] [p0574]  step: count(382), step_time 1404.33, mean_step_time 1435.28, it/s 0.7
  2211. [0130 17:39:10 @multigpu.py:323] [p0576]  step: count(389), step_time 1382.74, mean_step_time 1420.42, it/s 0.7
  2212. [0130 17:39:11 @multigpu.py:323] [p0115]  step: count(367), step_time 1404.09, mean_step_time 1430.14, it/s 0.7
  2213. [0130 17:39:11 @multigpu.py:323] [p0574]  step: count(383), step_time 1458.87, mean_step_time 1441.35, it/s 0.69
  2214. [0130 17:39:11 @multigpu.py:323] [p0576]  step: count(390), step_time 1433.6, mean_step_time 1421.77, it/s 0.7
  2215. [0130 17:39:13 @multigpu.py:323] [p0574]  step: count(384), step_time 1579.52, mean_step_time 1414.97, it/s 0.71
  2216. [0130 17:39:13 @multigpu.py:323] [p0576]  step: count(391), step_time 1378.32, mean_step_time 1415.01, it/s 0.71
  2217. [0130 17:39:13 @multigpu.py:323] [p0115]  step: count(368), step_time 1803.48, mean_step_time 1414.97, it/s 0.71
  2218. [0130 17:39:14 @multigpu.py:323] [p0115]  step: count(369), step_time 1356.61, mean_step_time 1411.38, it/s 0.71
  2219. [0130 17:39:14 @multigpu.py:323] [p0576]  step: count(392), step_time 1430.82, mean_step_time 1416.36, it/s 0.71
  2220. [0130 17:39:14 @multigpu.py:323] [p0574]  step: count(385), step_time 1488.43, mean_step_time 1420.69, it/s 0.7
  2221. [0130 17:39:15 @multigpu.py:323] [p0115]  step: count(370), step_time 1395.35, mean_step_time 1412.1, it/s 0.71
  2222. [0130 17:39:15 @multigpu.py:323] [p0576]  step: count(393), step_time 1378.79, mean_step_time 1412.24, it/s 0.71
  2223. [0130 17:39:15 @multigpu.py:323] [p0574]  step: count(386), step_time 1410.11, mean_step_time 1418.01, it/s 0.71
  2224. sending to address tcp://p0112:61216
  2225. ##### Sending to neptune:  online_score :  0.154189984997 , 1.0 #####
  2226. [u'online', 1.0]
  2227. receiving
  2228. sending to address tcp://p0112:61216
  2229. ##### Sending to neptune:  online_score :  0.154235092733 , 1.1 #####
  2230. [u'online', 1.1]
  2231. receiving
  2232. [0130 17:39:17 @multigpu.py:323] [p0115]  step: count(371), step_time 1412.42, mean_step_time 1413.89, it/s 0.71
  2233. [0130 17:39:17 @multigpu.py:323] [p0576]  step: count(394), step_time 1456.54, mean_step_time 1414.21, it/s 0.71
  2234. [0130 17:39:17 @multigpu.py:323] [p0574]  step: count(387), step_time 1381.87, mean_step_time 1415.86, it/s 0.71
  2235. [0130 17:39:18 @multigpu.py:323] [p0115]  step: count(372), step_time 1387.79, mean_step_time 1414.39, it/s 0.71
  2236. [0130 17:39:18 @multigpu.py:323] [p0574]  step: count(388), step_time 1394.24, mean_step_time 1417.53, it/s 0.71
  2237. [0130 17:39:18 @multigpu.py:323] [p0576]  step: count(395), step_time 1440.95, mean_step_time 1416.04, it/s 0.71
  2238. [0130 17:39:19 @multigpu.py:323] [p0115]  step: count(373), step_time 1417.49, mean_step_time 1415.01, it/s 0.71
  2239. [0130 17:39:20 @multigpu.py:323] [p0574]  step: count(389), step_time 1441.07, mean_step_time 1420.41, it/s 0.7
  2240. [0130 17:39:20 @multigpu.py:323] [p0576]  step: count(396), step_time 1436.47, mean_step_time 1414.45, it/s 0.71
  2241. [0130 17:39:21 @multigpu.py:323] [p0115]  step: count(374), step_time 1368.12, mean_step_time 1414.49, it/s 0.71
  2242. [0130 17:39:21 @multigpu.py:323] [p0574]  step: count(390), step_time 1398.35, mean_step_time 1421.53, it/s 0.7
  2243. [0130 17:39:21 @multigpu.py:323] [p0576]  step: count(397), step_time 1397.34, mean_step_time 1414.19, it/s 0.71
  2244. [0130 17:39:22 @multigpu.py:323] [p0115]  step: count(375), step_time 1369.1, mean_step_time 1411.72, it/s 0.71
  2245. [0130 17:39:22 @multigpu.py:323] [p0574]  step: count(391), step_time 1369.83, mean_step_time 1421.58, it/s 0.7
  2246. [0130 17:39:23 @multigpu.py:323] [p0576]  step: count(398), step_time 1483.03, mean_step_time 1416.42, it/s 0.71
  2247. [0130 17:39:24 @multigpu.py:323] [p0115]  step: count(376), step_time 1407.21, mean_step_time 1412.03, it/s 0.71
  2248. [0130 17:39:24 @multigpu.py:323] [p0574]  step: count(392), step_time 1344.19, mean_step_time 1413.44, it/s 0.71
  2249. [0130 17:39:24 @multigpu.py:323] [p0576]  step: count(399), step_time 1428.4, mean_step_time 1417.23, it/s 0.71
  2250. [0130 17:39:25 @multigpu.py:323] [p0115]  step: count(377), step_time 1352.72, mean_step_time 1409.7, it/s 0.71
  2251. [0130 17:39:25 @multigpu.py:323] [p0574]  step: count(393), step_time 1421.02, mean_step_time 1415.25, it/s 0.71
  2252. [0130 17:39:25 @multigpu.py:323] [p0576]  step: count(400), step_time 1398.51, mean_step_time 1418.24, it/s 0.71
  2253. sending debugging info...
  2254. sending to address tcp://p0112:61216
  2255. ##### Sending to neptune:  mean_delay :  0.156876291103 , 0.0 #####
  2256. ##### Sending to neptune:  max_delay :  0.156876291103 , -0.0 #####
  2257. sending to address tcp://p0112:61216
  2258. ##### Sending to neptune:  min_delay :  0.156876291103 , -0.0 #####
  2259. [u'delays', [0.0, -0.0, -0.0]]
  2260. receiving
  2261. ##### Sending to neptune:  cost :  0.156876804696 , -0.00845347996801 #####
  2262. sending to address tcp://p0112:61216
  2263. ##### Sending to neptune:  policy_loss :  0.156876804696 , 0.143131583929 #####
  2264. ##### Sending to neptune:  xentropy_loss :  0.156876804696 , -2.290143013 #####
  2265. ##### Sending to neptune:  value_loss :  0.156876804696 , 1.06496596336 #####
  2266. ##### Sending to neptune:  advantage :  0.156876804696 , -0.000597213045694 #####
  2267. ##### Sending to neptune:  pred_reward :  0.156876804696 , 0.332899808884 #####
  2268. ##### Sending to neptune:  max_logit :  0.156876804696 , 0.185626909137 #####
  2269. [u'loss', -0.00845347996801138, 0.1431315839290619, -2.2901430130004883, 1.0649659633636475, -0.000597213045693934, 0.332899808883667, 0.18562690913677216]
  2270. receiving
  2271. ##### Sending to neptune:  active_relus :  0.15687728107 , 8742840.19 #####
  2272. ##### Sending to neptune:  dp_per_s :  0.15687728107 , 89.5178593025 #####
  2273. [u'other', 8742840.19, 89.51785930253774]
  2274. receiving
  2275. [0130 17:39:26 @multigpu.py:323] [p0115]  step: count(378), step_time 1391.41, mean_step_time 1411.77, it/s 0.71
  2276. [0130 17:39:27 @multigpu.py:323] [p0574]  step: count(394), step_time 1364.77, mean_step_time 1410.26, it/s 0.71
  2277. [0130 17:39:27 @multigpu.py:323] [p0576]  step: count(401), step_time 1528.51, mean_step_time 1422.55, it/s 0.7
  2278. [0130 17:39:28 @multigpu.py:323] [p0115]  step: count(379), step_time 1352.05, mean_step_time 1408.9, it/s 0.71
  2279. [0130 17:39:28 @multigpu.py:323] [p0574]  step: count(395), step_time 1368.2, mean_step_time 1409.9, it/s 0.71
  2280. [0130 17:39:28 @multigpu.py:323] [p0576]  step: count(402), step_time 1385.77, mean_step_time 1421.99, it/s 0.7
  2281. [0130 17:39:29 @multigpu.py:323] [p0115]  step: count(380), step_time 1376.06, mean_step_time 1407.1, it/s 0.71
  2282. [0130 17:39:29 @multigpu.py:323] [p0574]  step: count(396), step_time 1420.37, mean_step_time 1410.23, it/s 0.71
  2283. [0130 17:39:30 @multigpu.py:323] [p0576]  step: count(403), step_time 1407.63, mean_step_time 1420.42, it/s 0.7
  2284. [0130 17:39:30 @multigpu.py:323] [p0115]  step: count(381), step_time 1404.98, mean_step_time 1405.86, it/s 0.71
  2285. [0130 17:39:31 @multigpu.py:323] [p0574]  step: count(397), step_time 1370.97, mean_step_time 1409.34, it/s 0.71
  2286. [0130 17:39:31 @multigpu.py:323] [p0576]  step: count(404), step_time 1454.24, mean_step_time 1423.36, it/s 0.7
  2287. [0130 17:39:32 @multigpu.py:323] [p0115]  step: count(382), step_time 1387.83, mean_step_time 1406.92, it/s 0.71
  2288. [0130 17:39:32 @multigpu.py:323] [p0574]  step: count(398), step_time 1396.96, mean_step_time 1410.16, it/s 0.71
  2289. [0130 17:39:33 @multigpu.py:323] [p0576]  step: count(405), step_time 1429.75, mean_step_time 1424.58, it/s 0.7
  2290. [0130 17:39:33 @multigpu.py:323] [p0115]  step: count(383), step_time 1429.27, mean_step_time 1406.86, it/s 0.71
  2291. [0130 17:39:34 @multigpu.py:323] [p0574]  step: count(399), step_time 1423.11, mean_step_time 1411.38, it/s 0.71
  2292. [0130 17:39:34 @multigpu.py:323] [p0576]  step: count(406), step_time 1410.93, mean_step_time 1422.87, it/s 0.7
  2293. [0130 17:39:35 @multigpu.py:323] [p0115]  step: count(384), step_time 1370.8, mean_step_time 1406.37, it/s 0.71
  2294. [0130 17:39:35 @multigpu.py:323] [p0574]  step: count(400), step_time 1392.73, mean_step_time 1411.12, it/s 0.71
  2295. sending debugging info...
  2296. sending to address tcp://p0112:61216
  2297. ##### Sending to neptune:  mean_delay :  0.159525871608 , 0.0 #####
  2298. sending to address tcp://p0112:61216
  2299. ##### Sending to neptune:  max_delay :  0.159525871608 , -0.0 #####
  2300. ##### Sending to neptune:  min_delay :  0.159525871608 , -0.0 #####
  2301. [u'delays', [0.0, -0.0, -0.0]]
  2302. ##### Sending to neptune:  active_workers :  0.159526103603 , 3 #####
  2303. receiving
  2304. ##### Sending to neptune:  cost :  0.159526361095 , -0.00946252513677 #####
  2305. sending to address tcp://p0112:61216
  2306. ##### Sending to neptune:  policy_loss :  0.159526361095 , 0.028890747577 #####
  2307. ##### Sending to neptune:  xentropy_loss :  0.159526361095 , -2.29028081894 #####
  2308. ##### Sending to neptune:  value_loss :  0.159526361095 , 1.05018687248 #####
  2309. ##### Sending to neptune:  advantage :  0.159526361095 , -0.000191484694369 #####
  2310. ##### Sending to neptune:  pred_reward :  0.159526361095 , 0.328690111637 #####
  2311. ##### Sending to neptune:  max_logit :  0.159526361095 , 0.185199126601 #####
  2312. [u'loss', -0.009462525136768818, 0.02889074757695198, -2.290280818939209, 1.0501868724822998, -0.0001914846943691373, 0.3286901116371155, 0.18519912660121918]
  2313. receiving
  2314. ##### Sending to neptune:  active_relus :  0.159526825216 , 8762594.07 #####
  2315. ##### Sending to neptune:  dp_per_s :  0.159526825216 , 90.186045923 #####
  2316. [u'other', 8762594.07, 90.18604592303011]
  2317. receiving
  2318. [0130 17:39:35 @multigpu.py:323] [p0576]  step: count(407), step_time 1431.07, mean_step_time 1424.57, it/s 0.7
  2319. sending to address tcp://p0112:61216
  2320. ##### Sending to neptune:  online_score :  0.15983586106 , 0.9 #####
  2321. [u'online', 0.9]
  2322. receiving
  2323. [0130 17:39:36 @multigpu.py:323] [p0115]  step: count(385), step_time 1370.4, mean_step_time 1406.4, it/s 0.71
  2324. [0130 17:39:36 @multigpu.py:323] [p0574]  step: count(401), step_time 1408.05, mean_step_time 1411.85, it/s 0.71
  2325. [0130 17:39:37 @multigpu.py:323] [p0576]  step: count(408), step_time 1494.54, mean_step_time 1429.4, it/s 0.7
  2326. [0130 17:39:38 @multigpu.py:323] [p0115]  step: count(386), step_time 1452.71, mean_step_time 1410.49, it/s 0.71
  2327. [0130 17:39:38 @multigpu.py:323] [p0574]  step: count(402), step_time 1369.83, mean_step_time 1410.12, it/s 0.71
  2328. [0130 17:39:38 @multigpu.py:323] [p0576]  step: count(409), step_time 1383.21, mean_step_time 1429.42, it/s 0.7
  2329. [0130 17:39:39 @multigpu.py:323] [p0115]  step: count(387), step_time 1360.87, mean_step_time 1408.33, it/s 0.71
  2330. [0130 17:39:39 @multigpu.py:323] [p0574]  step: count(403), step_time 1416.83, mean_step_time 1408.02, it/s 0.71
  2331. [0130 17:39:40 @multigpu.py:323] [p0576]  step: count(410), step_time 1437.49, mean_step_time 1429.62, it/s 0.7
  2332. [0130 17:39:41 @multigpu.py:323] [p0576]  step: count(411), step_time 1399.93, mean_step_time 1430.7, it/s 0.7
  2333. [0130 17:39:41 @multigpu.py:323] [p0574]  step: count(404), step_time 2033.89, mean_step_time 1430.74, it/s 0.7
  2334. [0130 17:39:41 @multigpu.py:323] [p0115]  step: count(388), step_time 2251.8, mean_step_time 1430.75, it/s 0.7
  2335. sending to address tcp://p0112:61216
  2336. ##### Sending to neptune:  online_score :  0.161397429109 , 1.8 #####
  2337. [u'online', 1.8]
  2338. receiving
  2339. [0130 17:39:43 @multigpu.py:323] [p0576]  step: count(412), step_time 1401.71, mean_step_time 1429.24, it/s 0.7
  2340. [0130 17:39:43 @multigpu.py:323] [p0115]  step: count(389), step_time 1446.82, mean_step_time 1435.26, it/s 0.7
  2341. [0130 17:39:43 @multigpu.py:323] [p0574]  step: count(405), step_time 1498.12, mean_step_time 1431.22, it/s 0.7
  2342. sending to address tcp://p0112:61216
  2343. ##### Sending to neptune:  online_score :  0.161699656049 , 1.6 #####
  2344. [u'online', 1.6]
  2345. receiving
  2346. [0130 17:39:44 @multigpu.py:323] [p0576]  step: count(413), step_time 1411.56, mean_step_time 1430.88, it/s 0.7
  2347. [0130 17:39:44 @multigpu.py:323] [p0115]  step: count(390), step_time 1435.59, mean_step_time 1437.27, it/s 0.7
  2348. [0130 17:39:44 @multigpu.py:323] [p0574]  step: count(406), step_time 1426.94, mean_step_time 1432.07, it/s 0.7
  2349. [0130 17:39:45 @multigpu.py:323] [p0576]  step: count(414), step_time 1422.95, mean_step_time 1429.2, it/s 0.7
  2350. [0130 17:39:45 @multigpu.py:323] [p0115]  step: count(391), step_time 1368.46, mean_step_time 1435.07, it/s 0.7
  2351. [0130 17:39:45 @multigpu.py:323] [p0574]  step: count(407), step_time 1412.92, mean_step_time 1433.62, it/s 0.7
  2352. [0130 17:39:47 @multigpu.py:323] [p0576]  step: count(415), step_time 1430.9, mean_step_time 1428.7, it/s 0.7
  2353. [0130 17:39:47 @multigpu.py:323] [p0115]  step: count(392), step_time 1425.59, mean_step_time 1436.96, it/s 0.7
  2354. [0130 17:39:47 @multigpu.py:323] [p0574]  step: count(408), step_time 1395.29, mean_step_time 1433.67, it/s 0.7
  2355. [0130 17:39:48 @multigpu.py:323] [p0115]  step: count(393), step_time 1358.57, mean_step_time 1434.02, it/s 0.7
  2356. [0130 17:39:48 @multigpu.py:323] [p0576]  step: count(416), step_time 1435.64, mean_step_time 1428.66, it/s 0.7
  2357. [0130 17:39:48 @multigpu.py:323] [p0574]  step: count(409), step_time 1427.05, mean_step_time 1432.97, it/s 0.7
  2358. [0130 17:39:50 @multigpu.py:323] [p0115]  step: count(394), step_time 1398.43, mean_step_time 1435.53, it/s 0.7
  2359. [0130 17:39:50 @multigpu.py:323] [p0576]  step: count(417), step_time 1410.86, mean_step_time 1429.33, it/s 0.7
  2360. [0130 17:39:50 @multigpu.py:323] [p0574]  step: count(410), step_time 1422.03, mean_step_time 1434.16, it/s 0.7
  2361. [0130 17:39:51 @multigpu.py:323] [p0115]  step: count(395), step_time 1386.76, mean_step_time 1436.42, it/s 0.7
  2362. [0130 17:39:51 @multigpu.py:323] [p0576]  step: count(418), step_time 1420.38, mean_step_time 1426.2, it/s 0.7
  2363. [0130 17:39:51 @multigpu.py:323] [p0574]  step: count(411), step_time 1371.37, mean_step_time 1434.23, it/s 0.7
  2364. [0130 17:39:52 @multigpu.py:323] [p0115]  step: count(396), step_time 1371.74, mean_step_time 1434.64, it/s 0.7
  2365. [0130 17:39:53 @multigpu.py:323] [p0574]  step: count(412), step_time 1423.33, mean_step_time 1438.19, it/s 0.7
  2366. [0130 17:39:53 @multigpu.py:323] [p0576]  step: count(419), step_time 1508.1, mean_step_time 1430.18, it/s 0.7
  2367. [0130 17:39:54 @multigpu.py:323] [p0115]  step: count(397), step_time 1408.93, mean_step_time 1437.45, it/s 0.7
  2368. [0130 17:39:54 @multigpu.py:323] [p0574]  step: count(413), step_time 1415.17, mean_step_time 1437.9, it/s 0.7
  2369. [0130 17:39:54 @multigpu.py:323] [p0576]  step: count(420), step_time 1399.22, mean_step_time 1430.22, it/s 0.7
  2370. [0130 17:39:55 @multigpu.py:323] [p0115]  step: count(398), step_time 1413.22, mean_step_time 1438.54, it/s 0.7
  2371. [0130 17:39:55 @multigpu.py:323] [p0574]  step: count(414), step_time 1410.68, mean_step_time 1440.19, it/s 0.69
  2372. [0130 17:39:56 @multigpu.py:323] [p0576]  step: count(421), step_time 1536.37, mean_step_time 1430.61, it/s 0.7
  2373. [0130 17:39:57 @multigpu.py:323] [p0115]  step: count(399), step_time 1444.12, mean_step_time 1443.15, it/s 0.69
  2374. [0130 17:39:57 @multigpu.py:323] [p0574]  step: count(415), step_time 1390.2, mean_step_time 1441.29, it/s 0.69
  2375. [0130 17:39:57 @multigpu.py:323] [p0576]  step: count(422), step_time 1412.54, mean_step_time 1431.95, it/s 0.7
  2376. [0130 17:39:58 @multigpu.py:323] [p0115]  step: count(400), step_time 1427.37, mean_step_time 1445.71, it/s 0.69
  2377. sending debugging info...
  2378. sending to address tcp://p0112:61216
  2379. ##### Sending to neptune:  mean_delay :  0.165948813889 , 0.0 #####
  2380. ##### Sending to neptune:  max_delay :  0.165948813889 , -0.0 #####
  2381. sending to address tcp://p0112:61216
  2382. ##### Sending to neptune:  min_delay :  0.165948813889 , -0.0 #####
  2383. [u'delays', [0.0, -0.0, -0.0]]
  2384. receiving
  2385. ##### Sending to neptune:  cost :  0.165949346357 , -0.00503255799413 #####
  2386. ##### Sending to neptune:  policy_loss :  0.165949346357 , 0.491833776236 #####
  2387. sending to address tcp://p0112:61216
  2388. ##### Sending to neptune:  xentropy_loss :  0.165949346357 , -2.29054236412 #####
  2389. ##### Sending to neptune:  value_loss :  0.165949346357 , 1.15454125404 #####
  2390. ##### Sending to neptune:  advantage :  0.165949346357 , -0.00217221933417 #####
  2391. ##### Sending to neptune:  pred_reward :  0.165949346357 , 0.325955986977 #####
  2392. ##### Sending to neptune:  max_logit :  0.165949346357 , 0.184464037418 #####
  2393. [u'loss', -0.0050325579941272736, 0.49183377623558044, -2.2905423641204834, 1.154541254043579, -0.0021722193341702223, 0.32595598697662354, 0.18446403741836548]
  2394. receiving
  2395. ##### Sending to neptune:  active_relus :  0.165949839685 , 8744566.58 #####
  2396. ##### Sending to neptune:  dp_per_s :  0.165949839685 , 89.9175458849 #####
  2397. [u'other', 8744566.58, 89.91754588494861]
  2398. receiving
  2399. [0130 17:39:58 @multigpu.py:323] [p0574]  step: count(416), step_time 1385.62, mean_step_time 1439.55, it/s 0.69
  2400. [0130 17:39:58 @multigpu.py:323] [p0576]  step: count(423), step_time 1421.15, mean_step_time 1432.63, it/s 0.7
  2401. [0130 17:39:59 @multigpu.py:323] [p0115]  step: count(401), step_time 1441.04, mean_step_time 1447.52, it/s 0.69
  2402. [0130 17:39:59 @multigpu.py:323] [p0574]  step: count(417), step_time 1377.82, mean_step_time 1439.9, it/s 0.69
  2403. [0130 17:40:00 @multigpu.py:323] [p0576]  step: count(424), step_time 1440.47, mean_step_time 1431.94, it/s 0.7
  2404. [0130 17:40:01 @multigpu.py:323] [p0115]  step: count(402), step_time 1358.7, mean_step_time 1446.06, it/s 0.69
  2405. [0130 17:40:01 @multigpu.py:323] [p0574]  step: count(418), step_time 1439.77, mean_step_time 1442.04, it/s 0.69
  2406. [0130 17:40:01 @multigpu.py:323] [p0576]  step: count(425), step_time 1406.44, mean_step_time 1430.77, it/s 0.7
  2407. [0130 17:40:02 @multigpu.py:323] [p0115]  step: count(403), step_time 1423.27, mean_step_time 1445.76, it/s 0.69
  2408. [0130 17:40:02 @multigpu.py:323] [p0574]  step: count(419), step_time 1406.12, mean_step_time 1441.19, it/s 0.69
  2409. [0130 17:40:03 @multigpu.py:323] [p0576]  step: count(426), step_time 1398.42, mean_step_time 1430.15, it/s 0.7
  2410. [0130 17:40:04 @multigpu.py:323] [p0115]  step: count(404), step_time 1408.34, mean_step_time 1447.64, it/s 0.69
  2411. [0130 17:40:04 @multigpu.py:323] [p0574]  step: count(420), step_time 1404.68, mean_step_time 1441.79, it/s 0.69
  2412. [0130 17:40:04 @multigpu.py:323] [p0576]  step: count(427), step_time 1447.22, mean_step_time 1430.95, it/s 0.7
  2413. [0130 17:40:05 @multigpu.py:323] [p0115]  step: count(405), step_time 1429.56, mean_step_time 1450.6, it/s 0.69
  2414. [0130 17:40:05 @multigpu.py:323] [p0574]  step: count(421), step_time 1381.4, mean_step_time 1440.45, it/s 0.69
  2415. [0130 17:40:05 @multigpu.py:323] [p0576]  step: count(428), step_time 1413.83, mean_step_time 1426.92, it/s 0.7
  2416. [0130 17:40:06 @multigpu.py:323] [p0115]  step: count(406), step_time 1383.02, mean_step_time 1447.11, it/s 0.69
  2417. [0130 17:40:07 @multigpu.py:323] [p0574]  step: count(422), step_time 1432.03, mean_step_time 1443.56, it/s 0.69
  2418. [0130 17:40:07 @multigpu.py:323] [p0576]  step: count(429), step_time 1435.3, mean_step_time 1429.52, it/s 0.7
  2419. [0130 17:40:08 @multigpu.py:323] [p0115]  step: count(407), step_time 1403.32, mean_step_time 1449.23, it/s 0.69
  2420. [0130 17:40:08 @multigpu.py:323] [p0574]  step: count(423), step_time 1403.98, mean_step_time 1442.92, it/s 0.69
  2421.  42%|####2     |423/1000[1[0130 17:40:08 @multigpu.py:323] [p0576]  step: count(430), step_time 1396.29, mean_step_time 1427.46, it/s 0.7
  2422. sending to address tcp://p0112:61216
  2423. ##### Sending to neptune:  online_score :  0.168805354701 , 1.6 #####
  2424. [u'online', 1.6]
  2425. receiving
  2426. sending to address tcp://p0112:61216
  2427. ##### Sending to neptune:  online_score :  0.168823744721 , 1.2 #####
  2428. [u'online', 1.2]
  2429. receiving
  2430. [0130 17:40:10 @multigpu.py:323] [p0576]  step: count(431), step_time 1448.01, mean_step_time 1429.87, it/s 0.7
  2431.  43%|####3     |431/1000[1[0130 17:40:10 @multigpu.py:323] [p0574]  step: count(424), step_time 1774.42, mean_step_time 1429.95, it/s 0.7
  2432. [0130 17:40:10 @multigpu.py:323] [p0115]  step: count(408), step_time 1866.32, mean_step_time 1429.96, it/s 0.7
  2433. sending to address tcp://p0112:61216
  2434. ##### Sending to neptune:  online_score :  0.169260831343 , 1.0 #####
  2435. [u'online', 1.0]
  2436. receiving
  2437. [0130 17:40:11 @multigpu.py:323] [p0576]  step: count(432), step_time 1369.59, mean_step_time 1428.26, it/s 0.7
  2438. [0130 17:40:11 @multigpu.py:323] [p0574]  step: count(425), step_time 1395.36, mean_step_time 1424.81, it/s 0.7
  2439. [0130 17:40:11 @multigpu.py:323] [p0115]  step: count(409), step_time 1451.41, mean_step_time 1430.19, it/s 0.7
  2440. [0130 17:40:13 @multigpu.py:323] [p0574]  step: count(426), step_time 1415.79, mean_step_time 1424.25, it/s 0.7
  2441. [0130 17:40:13 @multigpu.py:323] [p0115]  step: count(410), step_time 1391.93, mean_step_time 1428.01, it/s 0.7
  2442. [0130 17:40:13 @multigpu.py:323] [p0576]  step: count(433), step_time 1563.09, mean_step_time 1435.84, it/s 0.7
  2443. [0130 17:40:14 @multigpu.py:323] [p0115]  step: count(411), step_time 1372.63, mean_step_time 1428.21, it/s 0.7
  2444. [0130 17:40:14 @multigpu.py:323] [p0574]  step: count(427), step_time 1437.07, mean_step_time 1425.46, it/s 0.7
  2445. [0130 17:40:14 @multigpu.py:323] [p0576]  step: count(434), step_time 1453.46, mean_step_time 1437.36, it/s 0.7
  2446. [0130 17:40:15 @multigpu.py:323] [p0115]  step: count(412), step_time 1360.42, mean_step_time 1424.96, it/s 0.7
  2447. [0130 17:40:15 @multigpu.py:323] [p0574]  step: count(428), step_time 1463.36, mean_step_time 1428.86, it/s 0.7
  2448. [0130 17:40:16 @multigpu.py:323] [p0576]  step: count(435), step_time 1442.92, mean_step_time 1437.97, it/s 0.7
  2449. [0130 17:40:17 @multigpu.py:323] [p0115]  step: count(413), step_time 1390.98, mean_step_time 1426.58, it/s 0.7
  2450. [0130 17:40:17 @multigpu.py:323] [p0574]  step: count(429), step_time 1381.28, mean_step_time 1426.57, it/s 0.7
  2451. [0130 17:40:17 @multigpu.py:323] [p0576]  step: count(436), step_time 1397.1, mean_step_time 1436.04, it/s 0.7
  2452. [0130 17:40:18 @multigpu.py:323] [p0115]  step: count(414), step_time 1367.71, mean_step_time 1425.04, it/s 0.7
  2453. [0130 17:40:18 @multigpu.py:323] [p0574]  step: count(430), step_time 1453.28, mean_step_time 1428.14, it/s 0.7
  2454. [0130 17:40:18 @multigpu.py:323] [p0576]  step: count(437), step_time 1508.3, mean_step_time 1440.91, it/s 0.69
  2455. [0130 17:40:20 @multigpu.py:323] [p0115]  step: count(415), step_time 1425.78, mean_step_time 1426.99, it/s 0.7
  2456. [0130 17:40:20 @multigpu.py:323] [p0574]  step: count(431), step_time 1398.99, mean_step_time 1429.52, it/s 0.7
  2457. [0130 17:40:20 @multigpu.py:323] [p0576]  step: count(438), step_time 1407.49, mean_step_time 1440.27, it/s 0.69
  2458. [0130 17:40:21 @multigpu.py:323] [p0115]  step: count(416), step_time 1336.99, mean_step_time 1425.25, it/s 0.7
  2459. [0130 17:40:21 @multigpu.py:323] [p0574]  step: count(432), step_time 1432.96, mean_step_time 1430.0, it/s 0.7
  2460. [0130 17:40:21 @multigpu.py:323] [p0576]  step: count(439), step_time 1407.52, mean_step_time 1435.24, it/s 0.7
  2461. [0130 17:40:22 @multigpu.py:323] [p0115]  step: count(417), step_time 1408.77, mean_step_time 1425.25, it/s 0.7
  2462. [0130 17:40:23 @multigpu.py:323] [p0574]  step: count(433), step_time 1437.57, mean_step_time 1431.12, it/s 0.7
  2463. [0130 17:40:23 @multigpu.py:323] [p0576]  step: count(440), step_time 1469.58, mean_step_time 1438.75, it/s 0.7
  2464. [0130 17:40:24 @multigpu.py:323] [p0115]  step: count(418), step_time 1355.93, mean_step_time 1422.38, it/s 0.7
  2465. [0130 17:40:24 @multigpu.py:323] [p0574]  step: count(434), step_time 1321.66, mean_step_time 1426.67, it/s 0.7
  2466. [0130 17:40:24 @multigpu.py:323] [p0576]  step: count(441), step_time 1463.11, mean_step_time 1435.09, it/s 0.7
  2467. [0130 17:40:25 @multigpu.py:323] [p0115]  step: count(419), step_time 1403.54, mean_step_time 1420.35, it/s 0.7
  2468. [0130 17:40:25 @multigpu.py:323] [p0574]  step: count(435), step_time 1510.72, mean_step_time 1432.69, it/s 0.7
  2469. [0130 17:40:26 @multigpu.py:323] [p0576]  step: count(442), step_time 1466.47, mean_step_time 1437.79, it/s 0.7
  2470. [0130 17:40:26 @multigpu.py:323] [p0115]  step: count(420), step_time 1437.86, mean_step_time 1420.88, it/s 0.7
  2471. [0130 17:40:27 @multigpu.py:323] [p0574]  step: count(436), step_time 1422.17, mean_step_time 1434.52, it/s 0.7
  2472. [0130 17:40:27 @multigpu.py:323] [p0576]  step: count(443), step_time 1403.78, mean_step_time 1436.92, it/s 0.7
  2473. [0130 17:40:28 @multigpu.py:323] [p0115]  step: count(421), step_time 1481.53, mean_step_time 1422.9, it/s 0.7
  2474. [0130 17:40:28 @multigpu.py:323] [p0574]  step: count(437), step_time 1404.15, mean_step_time 1435.84, it/s 0.7
  2475. [0130 17:40:29 @multigpu.py:323] [p0576]  step: count(444), step_time 1448.02, mean_step_time 1437.3, it/s 0.7
  2476. [0130 17:40:29 @multigpu.py:323] [p0115]  step: count(422), step_time 1408.9, mean_step_time 1425.41, it/s 0.7
  2477. [0130 17:40:30 @multigpu.py:323] [p0574]  step: count(438), step_time 1427.92, mean_step_time 1435.25, it/s 0.7
  2478. [0130 17:40:30 @multigpu.py:323] [p0576]  step: count(445), step_time 1463.08, mean_step_time 1440.13, it/s 0.69
  2479. [0130 17:40:31 @multigpu.py:323] [p0115]  step: count(423), step_time 1399.13, mean_step_time 1424.2, it/s 0.7
  2480. [0130 17:40:31 @multigpu.py:323] [p0574]  step: count(439), step_time 1418.53, mean_step_time 1435.87, it/s 0.7
  2481. [0130 17:40:31 @multigpu.py:323] [p0576]  step: count(446), step_time 1432.68, mean_step_time 1441.84, it/s 0.69
  2482. [0130 17:40:32 @multigpu.py:323] [p0115]  step: count(424), step_time 1430.48, mean_step_time 1425.31, it/s 0.7
  2483. [0130 17:40:32 @multigpu.py:323] [p0574]  step: count(440), step_time 1376.25, mean_step_time 1434.45, it/s 0.7
  2484. [0130 17:40:33 @multigpu.py:323] [p0576]  step: count(447), step_time 1395.59, mean_step_time 1439.26, it/s 0.69
  2485. sending to address tcp://p0112:61216
  2486. ##### Sending to neptune:  online_score :  0.175806029704 , 2.3 #####
  2487. [u'online', 2.3]
  2488. receiving
  2489. [0130 17:40:34 @multigpu.py:323] [p0115]  step: count(425), step_time 1367.57, mean_step_time 1422.21, it/s 0.7
  2490. [0130 17:40:34 @multigpu.py:323] [p0574]  step: count(441), step_time 1522.2, mean_step_time 1441.49, it/s 0.69
  2491. [0130 17:40:34 @multigpu.py:323] [p0576]  step: count(448), step_time 1427.95, mean_step_time 1439.97, it/s 0.69
  2492. [0130 17:40:35 @multigpu.py:323] [p0115]  step: count(426), step_time 1404.57, mean_step_time 1423.29, it/s 0.7
  2493. [0130 17:40:35 @multigpu.py:323] [p0574]  step: count(442), step_time 1472.0, mean_step_time 1443.48, it/s 0.69
  2494. [0130 17:40:36 @multigpu.py:323] [p0576]  step: count(449), step_time 1448.3, mean_step_time 1440.62, it/s 0.69
  2495. [0130 17:40:36 @multigpu.py:323] [p0115]  step: count(427), step_time 1345.11, mean_step_time 1420.38, it/s 0.7
  2496. [0130 17:40:37 @multigpu.py:323] [p0574]  step: count(443), step_time 1446.05, mean_step_time 1445.59, it/s 0.69
  2497. [0130 17:40:37 @multigpu.py:323] [p0576]  step: count(450), step_time 1419.29, mean_step_time 1441.77, it/s 0.69
  2498. sending to address tcp://p0112:61216
  2499. ##### Sending to neptune:  online_score :  0.176930492719 , 1.4 #####
  2500. [u'online', 1.4]
  2501. ##### Sending to neptune:  active_workers :  0.176930545833 , 3 #####
  2502. receiving
  2503. [0130 17:40:39 @multigpu.py:323] [p0576]  step: count(451), step_time 1455.66, mean_step_time 1442.15, it/s 0.69
  2504. [0130 17:40:39 @multigpu.py:323] [p0115]  step: count(428), step_time 2300.21, mean_step_time 1442.07, it/s 0.69
  2505. [0130 17:40:39 @multigpu.py:323] [p0574]  step: count(444), step_time 1705.61, mean_step_time 1442.15, it/s 0.69
  2506.  43%|####2     |428/1000[1[0130 17:40:40 @multigpu.py:323] [p0115]  step: count(429), step_time 1387.77, mean_step_time 1438.89, it/s 0.69
  2507. [0130 17:40:40 @multigpu.py:323] [p0574]  step: count(445), step_time 1435.97, mean_step_time 1444.18, it/s 0.69
  2508. [0130 17:40:40 @multigpu.py:323] [p0576]  step: count(452), step_time 1460.36, mean_step_time 1446.69, it/s 0.69
  2509. sending to address tcp://p0112:61216
  2510. ##### Sending to neptune:  online_score :  0.177810673846 , 1.2 #####
  2511. [u'online', 1.2]
  2512. receiving
  2513. [0130 17:40:41 @multigpu.py:323] [p0574]  step: count(446), step_time 1355.41, mean_step_time 1441.16, it/s 0.69
  2514. [0130 17:40:41 @multigpu.py:323] [p0115]  step: count(430), step_time 1422.15, mean_step_time 1440.4, it/s 0.69
  2515. [0130 17:40:42 @multigpu.py:323] [p0576]  step: count(453), step_time 1466.27, mean_step_time 1441.85, it/s 0.69
  2516. [0130 17:40:43 @multigpu.py:323] [p0574]  step: count(447), step_time 1443.97, mean_step_time 1441.5, it/s 0.69
  2517. [0130 17:40:43 @multigpu.py:323] [p0115]  step: count(431), step_time 1449.53, mean_step_time 1444.25, it/s 0.69
  2518. [0130 17:40:43 @multigpu.py:323] [p0576]  step: count(454), step_time 1419.22, mean_step_time 1440.13, it/s 0.69
  2519. [0130 17:40:44 @multigpu.py:323] [p0574]  step: count(448), step_time 1394.73, mean_step_time 1438.07, it/s 0.7
  2520. [0130 17:40:44 @multigpu.py:323] [p0115]  step: count(432), step_time 1402.93, mean_step_time 1446.37, it/s 0.69
  2521. [0130 17:40:44 @multigpu.py:323] [p0576]  step: count(455), step_time 1422.21, mean_step_time 1439.1, it/s 0.69
  2522. [0130 17:40:46 @multigpu.py:323] [p0115]  step: count(433), step_time 1391.4, mean_step_time 1446.39, it/s 0.69
  2523. [0130 17:40:46 @multigpu.py:323] [p0574]  step: count(449), step_time 1448.14, mean_step_time 1441.42, it/s 0.69
  2524. [0130 17:40:46 @multigpu.py:323] [p0576]  step: count(456), step_time 1384.99, mean_step_time 1438.49, it/s 0.7
  2525. [0130 17:40:47 @multigpu.py:323] [p0115]  step: count(434), step_time 1393.72, mean_step_time 1447.69, it/s 0.69
  2526. [0130 17:40:47 @multigpu.py:323] [p0574]  step: count(450), step_time 1379.57, mean_step_time 1437.73, it/s 0.7
  2527. [0130 17:40:47 @multigpu.py:323] [p0576]  step: count(457), step_time 1415.17, mean_step_time 1433.84, it/s 0.7
  2528. [0130 17:40:48 @multigpu.py:323] [p0574]  step: count(451), step_time 1423.27, mean_step_time 1438.94, it/s 0.69
  2529. [0130 17:40:48 @multigpu.py:323] [p0115]  step: count(435), step_time 1451.37, mean_step_time 1448.97, it/s 0.69
  2530. [0130 17:40:49 @multigpu.py:323] [p0576]  step: count(458), step_time 1438.82, mean_step_time 1435.4, it/s 0.7
  2531. [0130 17:40:50 @multigpu.py:323] [p0574]  step: count(452), step_time 1390.49, mean_step_time 1436.82, it/s 0.7
  2532. [0130 17:40:50 @multigpu.py:323] [p0115]  step: count(436), step_time 1378.98, mean_step_time 1451.07, it/s 0.69
  2533. [0130 17:40:50 @multigpu.py:323] [p0576]  step: count(459), step_time 1444.31, mean_step_time 1437.24, it/s 0.7
  2534. [0130 17:40:51 @multigpu.py:323] [p0574]  step: count(453), step_time 1400.46, mean_step_time 1434.96, it/s 0.7
  2535. [0130 17:40:51 @multigpu.py:323] [p0115]  step: count(437), step_time 1464.35, mean_step_time 1453.85, it/s 0.69
  2536. [0130 17:40:51 @multigpu.py:323] [p0576]  step: count(460), step_time 1397.47, mean_step_time 1433.64, it/s 0.7
  2537. [0130 17:40:53 @multigpu.py:323] [p0574]  step: count(454), step_time 1389.83, mean_step_time 1438.37, it/s 0.7
  2538. [0130 17:40:53 @multigpu.py:323] [p0115]  step: count(438), step_time 1365.71, mean_step_time 1454.34, it/s 0.69
  2539. [0130 17:40:53 @multigpu.py:323] [p0576]  step: count(461), step_time 1436.6, mean_step_time 1432.31, it/s 0.7
  2540. [0130 17:40:54 @multigpu.py:323] [p0574]  step: count(455), step_time 1378.02, mean_step_time 1431.74, it/s 0.7
  2541. [0130 17:40:54 @multigpu.py:323] [p0115]  step: count(439), step_time 1415.1, mean_step_time 1454.92, it/s 0.69
  2542. [0130 17:40:54 @multigpu.py:323] [p0576]  step: count(462), step_time 1447.98, mean_step_time 1431.39, it/s 0.7
  2543. [0130 17:40:55 @multigpu.py:323] [p0574]  step: count(456), step_time 1394.3, mean_step_time 1430.34, it/s 0.7
  2544. [0130 17:40:56 @multigpu.py:323] [p0115]  step: count(440), step_time 1380.95, mean_step_time 1452.07, it/s 0.69
  2545. [0130 17:40:56 @multigpu.py:323] [p0576]  step: count(463), step_time 1442.55, mean_step_time 1433.33, it/s 0.7
  2546. [0130 17:40:57 @multigpu.py:323] [p0574]  step: count(457), step_time 1398.26, mean_step_time 1430.05, it/s 0.7
  2547. [0130 17:40:57 @multigpu.py:323] [p0115]  step: count(441), step_time 1458.1, mean_step_time 1450.9, it/s 0.69
  2548. [0130 17:40:57 @multigpu.py:323] [p0576]  step: count(464), step_time 1360.36, mean_step_time 1428.94, it/s 0.7
  2549. [0130 17:40:58 @multigpu.py:323] [p0574]  step: count(458), step_time 1472.43, mean_step_time 1432.27, it/s 0.7
  2550. [0130 17:40:58 @multigpu.py:323] [p0115]  step: count(442), step_time 1427.88, mean_step_time 1451.85, it/s 0.69
  2551. [0130 17:40:59 @multigpu.py:323] [p0576]  step: count(465), step_time 1417.51, mean_step_time 1426.66, it/s 0.7
  2552. [0130 17:41:00 @multigpu.py:323] [p0574]  step: count(459), step_time 1415.56, mean_step_time 1432.13, it/s 0.7
  2553. [0130 17:41:00 @multigpu.py:323] [p0115]  step: count(443), step_time 1447.21, mean_step_time 1454.26, it/s 0.69
  2554. [0130 17:41:00 @multigpu.py:323] [p0576]  step: count(466), step_time 1473.78, mean_step_time 1428.72, it/s 0.7
  2555. [0130 17:41:01 @multigpu.py:323] [p0574]  step: count(460), step_time 1463.47, mean_step_time 1436.49, it/s 0.7
  2556. [0130 17:41:01 @multigpu.py:323] [p0115]  step: count(444), step_time 1474.85, mean_step_time 1456.47, it/s 0.69
  2557. [0130 17:41:01 @multigpu.py:323] [p0576]  step: count(467), step_time 1465.99, mean_step_time 1432.24, it/s 0.7
  2558. [0130 17:41:03 @multigpu.py:323] [p0574]  step: count(461), step_time 1431.21, mean_step_time 1431.94, it/s 0.7
  2559. [0130 17:41:03 @multigpu.py:323] [p0115]  step: count(445), step_time 1452.57, mean_step_time 1460.72, it/s 0.68
  2560. [0130 17:41:03 @multigpu.py:323] [p0576]  step: count(468), step_time 1414.81, mean_step_time 1431.58, it/s 0.7
  2561. sending to address tcp://p0112:61216
  2562. ##### Sending to neptune:  online_score :  0.184176382754 , 1.5 #####
  2563. [u'online', 1.5]
  2564. receiving
  2565. [0130 17:41:04 @multigpu.py:323] [p0574]  step: count(462), step_time 1385.28, mean_step_time 1427.6, it/s 0.7
  2566. [0130 17:41:04 @multigpu.py:323] [p0115]  step: count(446), step_time 1353.37, mean_step_time 1458.16, it/s 0.69
  2567. [0130 17:41:04 @multigpu.py:323] [p0576]  step: count(469), step_time 1445.93, mean_step_time 1431.46, it/s 0.7
  2568. [0130 17:41:05 @multigpu.py:323] [p0574]  step: count(463), step_time 1400.82, mean_step_time 1425.34, it/s 0.7
  2569. [0130 17:41:06 @multigpu.py:323] [p0115]  step: count(447), step_time 1450.28, mean_step_time 1463.42, it/s 0.68
  2570. [0130 17:41:06 @multigpu.py:323] [p0576]  step: count(470), step_time 1445.53, mean_step_time 1432.78, it/s 0.7
  2571. [0130 17:41:07 @multigpu.py:323] [p0576]  step: count(471), step_time 1439.44, mean_step_time 1431.97, it/s 0.7
  2572. [0130 17:41:07 @multigpu.py:323] [p0115]  step: count(448), step_time 1671.13, mean_step_time 1431.97, it/s 0.7
  2573. [0130 17:41:07 @multigpu.py:323] [p0574]  step: count(464), step_time 1838.23, mean_step_time 1431.97, it/s 0.7
  2574. [0130 17:41:09 @multigpu.py:323] [p0576]  step: count(472), step_time 1414.98, mean_step_time 1429.7, it/s 0.7
  2575. [0130 17:41:09 @multigpu.py:323] [p0574]  step: count(465), step_time 1447.22, mean_step_time 1432.53, it/s 0.7
  2576. [0130 17:41:09 @multigpu.py:323] [p0115]  step: count(449), step_time 1450.3, mean_step_time 1435.09, it/s 0.7
  2577. sending to address tcp://p0112:61216
  2578. ##### Sending to neptune:  online_score :  0.185665526655 , 1.7 #####
  2579. [u'online', 1.7]
  2580. receiving
  2581. sending to address tcp://p0112:61216
  2582. ##### Sending to neptune:  online_score :  0.185782292485 , 1.8 #####
  2583. [u'online', 1.8]
  2584. receiving
  2585. [0130 17:41:10 @multigpu.py:323] [p0115]  step: count(450), step_time 1355.29, mean_step_time 1431.75, it/s 0.7
  2586. [0130 17:41:10 @multigpu.py:323] [p0574]  step: count(466), step_time 1452.54, mean_step_time 1437.39, it/s 0.7
  2587. [0130 17:41:10 @multigpu.py:323] [p0576]  step: count(473), step_time 1488.28, mean_step_time 1430.8, it/s 0.7
  2588. [0130 17:41:12 @multigpu.py:323] [p0115]  step: count(451), step_time 1474.42, mean_step_time 1433.0, it/s 0.7
  2589. [0130 17:41:12 @multigpu.py:323] [p0574]  step: count(467), step_time 1381.35, mean_step_time 1434.26, it/s 0.7
  2590. [0130 17:41:12 @multigpu.py:323] [p0576]  step: count(474), step_time 1435.59, mean_step_time 1431.62, it/s 0.7
  2591. [0130 17:41:13 @multigpu.py:323] [p0574]  step: count(468), step_time 1443.75, mean_step_time 1436.71, it/s 0.7
  2592. [0130 17:41:13 @multigpu.py:323] [p0115]  step: count(452), step_time 1461.72, mean_step_time 1435.93, it/s 0.7
  2593. [0130 17:41:13 @multigpu.py:323] [p0576]  step: count(475), step_time 1461.87, mean_step_time 1433.6, it/s 0.7
  2594. [0130 17:41:14 @multigpu.py:323] [p0574]  step: count(469), step_time 1376.0, mean_step_time 1433.1, it/s 0.7
  2595. [0130 17:41:14 @multigpu.py:323] [p0115]  step: count(453), step_time 1415.55, mean_step_time 1437.14, it/s 0.7
  2596. [0130 17:41:14 @multigpu.py:323] [p0576]  step: count(476), step_time 1399.07, mean_step_time 1434.3, it/s 0.7
  2597. [0130 17:41:16 @multigpu.py:323] [p0574]  step: count(470), step_time 1423.37, mean_step_time 1435.29, it/s 0.7
  2598. [0130 17:41:16 @multigpu.py:323] [p0115]  step: count(454), step_time 1441.31, mean_step_time 1439.52, it/s 0.69
  2599. [0130 17:41:16 @multigpu.py:323] [p0576]  step: count(477), step_time 1432.48, mean_step_time 1435.17, it/s 0.7
  2600. [0130 17:41:17 @multigpu.py:323] [p0574]  step: count(471), step_time 1408.02, mean_step_time 1434.53, it/s 0.7
  2601. [0130 17:41:17 @multigpu.py:323] [p0115]  step: count(455), step_time 1413.35, mean_step_time 1437.62, it/s 0.7
  2602. [0130 17:41:17 @multigpu.py:323] [p0576]  step: count(478), step_time 1450.45, mean_step_time 1435.75, it/s 0.7
  2603. [0130 17:41:19 @multigpu.py:323] [p0574]  step: count(472), step_time 1383.97, mean_step_time 1434.2, it/s 0.7
  2604. [0130 17:41:19 @multigpu.py:323] [p0115]  step: count(456), step_time 1441.25, mean_step_time 1440.73, it/s 0.69
  2605. [0130 17:41:19 @multigpu.py:323] [p0576]  step: count(479), step_time 1452.88, mean_step_time 1436.18, it/s 0.7
  2606. [0130 17:41:20 @multigpu.py:323] [p0574]  step: count(473), step_time 1496.55, mean_step_time 1439.01, it/s 0.69
  2607. [0130 17:41:20 @multigpu.py:323] [p0115]  step: count(457), step_time 1428.56, mean_step_time 1438.94, it/s 0.69
  2608. [0130 17:41:20 @multigpu.py:323] [p0576]  step: count(480), step_time 1485.43, mean_step_time 1440.58, it/s 0.69
  2609. [0130 17:41:22 @multigpu.py:323] [p0574]  step: count(474), step_time 1446.31, mean_step_time 1441.83, it/s 0.69
  2610. [0130 17:41:22 @multigpu.py:323] [p0115]  step: count(458), step_time 1387.74, mean_step_time 1440.05, it/s 0.69
  2611. [0130 17:41:22 @multigpu.py:323] [p0576]  step: count(481), step_time 1456.43, mean_step_time 1441.57, it/s 0.69
  2612. [0130 17:41:23 @multigpu.py:323] [p0574]  step: count(475), step_time 1402.51, mean_step_time 1443.06, it/s 0.69
  2613. [0130 17:41:23 @multigpu.py:323] [p0115]  step: count(459), step_time 1490.29, mean_step_time 1443.81, it/s 0.69
  2614. [0130 17:41:23 @multigpu.py:323] [p0576]  step: count(482), step_time 1437.77, mean_step_time 1441.06, it/s 0.69
  2615. [0130 17:41:24 @multigpu.py:323] [p0574]  step: count(476), step_time 1461.35, mean_step_time 1446.41, it/s 0.69
  2616. [0130 17:41:24 @multigpu.py:323] [p0115]  step: count(460), step_time 1413.76, mean_step_time 1445.45, it/s 0.69
  2617. [0130 17:41:25 @multigpu.py:323] [p0576]  step: count(483), step_time 1371.29, mean_step_time 1437.49, it/s 0.7
  2618. [0130 17:41:26 @multigpu.py:323] [p0574]  step: count(477), step_time 1460.21, mean_step_time 1449.51, it/s 0.69
  2619. [0130 17:41:26 @multigpu.py:323] [p0115]  step: count(461), step_time 1452.77, mean_step_time 1445.18, it/s 0.69
  2620. [0130 17:41:26 @multigpu.py:323] [p0576]  step: count(484), step_time 1469.12, mean_step_time 1442.93, it/s 0.69
  2621. [0130 17:41:27 @multigpu.py:323] [p0574]  step: count(478), step_time 1411.56, mean_step_time 1446.46, it/s 0.69
  2622. [0130 17:41:27 @multigpu.py:323] [p0115]  step: count(462), step_time 1413.91, mean_step_time 1444.48, it/s 0.69
  2623. [0130 17:41:27 @multigpu.py:323] [p0576]  step: count(485), step_time 1386.82, mean_step_time 1441.4, it/s 0.69
  2624. [0130 17:41:29 @multigpu.py:323] [p0574]  step: count(479), step_time 1431.48, mean_step_time 1447.26, it/s 0.69
  2625. [0130 17:41:29 @multigpu.py:323] [p0115]  step: count(463), step_time 1387.01, mean_step_time 1441.47, it/s 0.69
  2626. [0130 17:41:29 @multigpu.py:323] [p0576]  step: count(486), step_time 1379.58, mean_step_time 1436.69, it/s 0.7
  2627. [0130 17:41:30 @multigpu.py:323] [p0115]  step: count(464), step_time 1401.1, mean_step_time 1437.78, it/s 0.7
  2628. [0130 17:41:30 @multigpu.py:323] [p0574]  step: count(480), step_time 1439.94, mean_step_time 1446.08, it/s 0.69
  2629. [0130 17:41:30 @multigpu.py:323] [p0576]  step: count(487), step_time 1441.15, mean_step_time 1435.45, it/s 0.7
  2630. [0130 17:41:31 @multigpu.py:323] [p0115]  step: count(465), step_time 1415.79, mean_step_time 1435.95, it/s 0.7
  2631. [0130 17:41:32 @multigpu.py:323] [p0574]  step: count(481), step_time 1418.57, mean_step_time 1445.45, it/s 0.69
  2632. [0130 17:41:32 @multigpu.py:323] [p0576]  step: count(488), step_time 1418.84, mean_step_time 1435.65, it/s 0.7
  2633. [0130 17:41:33 @multigpu.py:323] [p0115]  step: count(466), step_time 1443.08, mean_step_time 1440.43, it/s 0.69
  2634. [0130 17:41:33 @multigpu.py:323] [p0574]  step: count(482), step_time 1420.62, mean_step_time 1447.22, it/s 0.69
  2635. [0130 17:41:33 @multigpu.py:323] [p0576]  step: count(489), step_time 1418.78, mean_step_time 1434.29, it/s 0.7
  2636. sending to address tcp://p0112:61216
  2637. ##### Sending to neptune:  online_score :  0.192464135223 , 1.2 #####
  2638. [u'online', 1.2]
  2639. receiving
  2640. [0130 17:41:34 @multigpu.py:323] [p0574]  step: count(483), step_time 1385.07, mean_step_time 1446.43, it/s 0.69
  2641. sending to address tcp://p0112:61216
  2642. ##### Sending to neptune:  online_score :  0.192708853881 , 1.7 #####
  2643. [u'online', 1.7]
  2644. receiving
  2645. [0130 17:41:34 @multigpu.py:323] [p0576]  step: count(490), step_time 1362.98, mean_step_time 1430.16, it/s 0.7
  2646. [0130 17:41:34 @multigpu.py:323] [p0115]  step: count(467), step_time 1487.79, mean_step_time 1442.31, it/s 0.69
  2647. [0130 17:41:36 @multigpu.py:323] [p0115]  step: count(468), step_time 1426.91, mean_step_time 1430.1, it/s 0.7
  2648. [0130 17:41:36 @multigpu.py:323] [p0574]  step: count(484), step_time 1515.46, mean_step_time 1430.29, it/s 0.7
  2649. [0130 17:41:36 @multigpu.py:323] [p0576]  step: count(491), step_time 1442.33, mean_step_time 1430.31, it/s 0.7
  2650. sending to address tcp://p0112:61216
  2651. ##### Sending to neptune:  online_score :  0.193346670535 , 1.6 #####
  2652. [u'online', 1.6]
  2653. receiving
  2654. [0130 17:41:37 @multigpu.py:323] [p0115]  step: count(469), step_time 1406.09, mean_step_time 1427.89, it/s 0.7
  2655. [0130 17:41:37 @multigpu.py:323] [p0574]  step: count(485), step_time 1420.3, mean_step_time 1428.95, it/s 0.7
  2656. [0130 17:41:37 @multigpu.py:323] [p0576]  step: count(492), step_time 1433.82, mean_step_time 1431.25, it/s 0.7
  2657. [0130 17:41:39 @multigpu.py:323] [p0115]  step: count(470), step_time 1469.25, mean_step_time 1433.58, it/s 0.7
  2658. [0130 17:41:39 @multigpu.py:323] [p0574]  step: count(486), step_time 1462.15, mean_step_time 1429.43, it/s 0.7
  2659. [0130 17:41:39 @multigpu.py:323] [p0576]  step: count(493), step_time 1465.1, mean_step_time 1430.09, it/s 0.7
  2660. [0130 17:41:40 @multigpu.py:323] [p0574]  step: count(487), step_time 1401.71, mean_step_time 1430.44, it/s 0.7
  2661. [0130 17:41:40 @multigpu.py:323] [p0115]  step: count(471), step_time 1413.02, mean_step_time 1430.51, it/s 0.7
  2662. [0130 17:41:40 @multigpu.py:323] [p0576]  step: count(494), step_time 1428.58, mean_step_time 1429.74, it/s 0.7
  2663. [0130 17:41:42 @multigpu.py:323] [p0574]  step: count(488), step_time 1397.83, mean_step_time 1428.15, it/s 0.7
  2664. [0130 17:41:42 @multigpu.py:323] [p0115]  step: count(472), step_time 1402.81, mean_step_time 1427.57, it/s 0.7
  2665. [0130 17:41:42 @multigpu.py:323] [p0576]  step: count(495), step_time 1393.27, mean_step_time 1426.31, it/s 0.7
  2666. [0130 17:41:43 @multigpu.py:323] [p0574]  step: count(489), step_time 1419.44, mean_step_time 1430.32, it/s 0.7
  2667. [0130 17:41:43 @multigpu.py:323] [p0576]  step: count(496), step_time 1402.95, mean_step_time 1426.5, it/s 0.7
  2668. [0130 17:41:43 @multigpu.py:323] [p0115]  step: count(473), step_time 1439.46, mean_step_time 1428.76, it/s 0.7
  2669. [0130 17:41:44 @multigpu.py:323] [p0574]  step: count(490), step_time 1411.47, mean_step_time 1429.73, it/s 0.7
  2670. [0130 17:41:44 @multigpu.py:323] [p0576]  step: count(497), step_time 1391.16, mean_step_time 1424.44, it/s 0.7
  2671. [0130 17:41:44 @multigpu.py:323] [p0115]  step: count(474), step_time 1432.71, mean_step_time 1428.33, it/s 0.7
  2672. [0130 17:41:46 @multigpu.py:323] [p0576]  step: count(498), step_time 1395.99, mean_step_time 1421.71, it/s 0.7
  2673. [0130 17:41:46 @multigpu.py:323] [p0574]  step: count(491), step_time 1425.6, mean_step_time 1430.6, it/s 0.7
  2674. [0130 17:41:46 @multigpu.py:323] [p0115]  step: count(475), step_time 1449.3, mean_step_time 1430.13, it/s 0.7
  2675. [0130 17:41:47 @multigpu.py:323] [p0574]  step: count(492), step_time 1396.91, mean_step_time 1431.25, it/s 0.7
  2676. [0130 17:41:47 @multigpu.py:323] [p0576]  step: count(499), step_time 1446.81, mean_step_time 1421.41, it/s 0.7
  2677. [0130 17:41:47 @multigpu.py:323] [p0115]  step: count(476), step_time 1454.84, mean_step_time 1430.81, it/s 0.7
  2678. [0130 17:41:49 @multigpu.py:323] [p0574]  step: count(493), step_time 1403.47, mean_step_time 1426.6, it/s 0.7
  2679. [0130 17:41:49 @multigpu.py:323] [p0576]  step: count(500), step_time 1426.58, mean_step_time 1418.47, it/s 0.7
  2680. sending debugging info...
  2681. sending to address tcp://p0112:61216
  2682. ##### Sending to neptune:  mean_delay :  0.19667889244 , 0.0 #####
  2683. ##### Sending to neptune:  max_delay :  0.19667889244 , -0.0 #####
  2684. sending to address tcp://p0112:61216
  2685. ##### Sending to neptune:  min_delay :  0.19667889244 , -0.0 #####
  2686. [u'delays', [0.0, -0.0, -0.0]]
  2687. ##### Sending to neptune:  active_workers :  0.196679146356 , 3 #####
  2688. receiving
  2689. ##### Sending to neptune:  cost :  0.196679431332 , -0.00439015356824 #####
  2690. ##### Sending to neptune:  policy_loss :  0.196679431332 , 0.453959643841 #####
  2691. sending to address tcp://p0112:61216
  2692. ##### Sending to neptune:  xentropy_loss :  0.196679431332 , -2.29136276245 #####
  2693. ##### Sending to neptune:  value_loss :  0.196679431332 , 1.27546346188 #####
  2694. ##### Sending to neptune:  advantage :  0.196679431332 , -0.00191708060447 #####
  2695. ##### Sending to neptune:  pred_reward :  0.196679431332 , 0.342677921057 #####
  2696. ##### Sending to neptune:  max_logit :  0.196679431332 , 0.181843400002 #####
  2697. [u'loss', -0.00439015356823802, 0.4539596438407898, -2.291362762451172, 1.2754634618759155, -0.0019170806044712663, 0.34267792105674744, 0.18184340000152588]
  2698. receiving
  2699. ##### Sending to neptune:  active_relus :  0.19667988552 , 8799167.4 #####
  2700. ##### Sending to neptune:  dp_per_s :  0.19667988552 , 89.3428788891 #####
  2701. [u'other', 8799167.4, 89.34287888906876]
  2702. receiving
  2703. [0130 17:41:49 @multigpu.py:323] [p0115]  step: count(477), step_time 1390.74, mean_step_time 1428.92, it/s 0.7
  2704. [0130 17:41:50 @multigpu.py:323] [p0574]  step: count(494), step_time 1432.89, mean_step_time 1425.93, it/s 0.7
  2705. [0130 17:41:50 @multigpu.py:323] [p0115]  step: count(478), step_time 1369.78, mean_step_time 1428.02, it/s 0.7
  2706. [0130 17:41:50 @multigpu.py:323] [p0576]  step: count(501), step_time 1493.99, mean_step_time 1420.34, it/s 0.7
  2707. [0130 17:41:51 @multigpu.py:323] [p0574]  step: count(495), step_time 1441.23, mean_step_time 1427.86, it/s 0.7
  2708. [0130 17:41:51 @multigpu.py:323] [p0115]  step: count(479), step_time 1400.71, mean_step_time 1423.54, it/s 0.7
  2709. [0130 17:41:52 @multigpu.py:323] [p0576]  step: count(502), step_time 1473.53, mean_step_time 1422.13, it/s 0.7
  2710. [0130 17:41:53 @multigpu.py:323] [p0574]  step: count(496), step_time 1426.57, mean_step_time 1426.12, it/s 0.7
  2711. [0130 17:41:53 @multigpu.py:323] [p0115]  step: count(480), step_time 1464.67, mean_step_time 1426.09, it/s 0.7
  2712. [0130 17:41:53 @multigpu.py:323] [p0576]  step: count(503), step_time 1417.0, mean_step_time 1424.42, it/s 0.7
  2713. [0130 17:41:54 @multigpu.py:323] [p0574]  step: count(497), step_time 1475.96, mean_step_time 1426.91, it/s 0.7
  2714. [0130 17:41:54 @multigpu.py:323] [p0115]  step: count(481), step_time 1456.84, mean_step_time 1426.29, it/s 0.7
  2715. [0130 17:41:54 @multigpu.py:323] [p0576]  step: count(504), step_time 1443.88, mean_step_time 1423.16, it/s 0.7
  2716. [0130 17:41:56 @multigpu.py:323] [p0574]  step: count(498), step_time 1423.72, mean_step_time 1427.52, it/s 0.7
  2717. [0130 17:41:56 @multigpu.py:323] [p0115]  step: count(482), step_time 1441.86, mean_step_time 1427.69, it/s 0.7
  2718. [0130 17:41:56 @multigpu.py:323] [p0576]  step: count(505), step_time 1450.04, mean_step_time 1426.32, it/s 0.7
  2719. [0130 17:41:57 @multigpu.py:323] [p0574]  step: count(499), step_time 1483.14, mean_step_time 1430.1, it/s 0.7
  2720. [0130 17:41:57 @multigpu.py:323] [p0115]  step: count(483), step_time 1466.56, mean_step_time 1431.67, it/s 0.7
  2721. [0130 17:41:57 @multigpu.py:323] [p0576]  step: count(506), step_time 1446.31, mean_step_time 1429.65, it/s 0.7
  2722. [0130 17:41:59 @multigpu.py:323] [p0115]  step: count(484), step_time 1380.52, mean_step_time 1430.64, it/s 0.7
  2723. [0130 17:41:59 @multigpu.py:323] [p0574]  step: count(500), step_time 1416.52, mean_step_time 1428.93, it/s 0.7
  2724. sending debugging info...
  2725. sending to address tcp://p0112:61216
  2726. ##### Sending to neptune:  mean_delay :  0.199472724663 , 0.0 #####
  2727. ##### Sending to neptune:  max_delay :  0.199472724663 , -0.0 #####
  2728. sending to address tcp://p0112:61216
  2729. ##### Sending to neptune:  min_delay :  0.199472724663 , -0.0 #####
  2730. [u'delays', [0.0, -0.0, -0.0]]
  2731. receiving
  2732. ##### Sending to neptune:  cost :  0.199473265807 , -0.00781606975943 #####
  2733. sending to address tcp://p0112:61216
  2734. ##### Sending to neptune:  policy_loss :  0.199473265807 , 0.0337311848998 #####
  2735. ##### Sending to neptune:  xentropy_loss :  0.199473265807 , -2.29143595695 #####
  2736. ##### Sending to neptune:  value_loss :  0.199473265807 , 1.2572479248 #####
  2737. ##### Sending to neptune:  advantage :  0.199473265807 , -3.6175846617e-05 #####
  2738. ##### Sending to neptune:  pred_reward :  0.199473265807 , 0.347458809614 #####
  2739. ##### Sending to neptune:  max_logit :  0.199473265807 , 0.181484788656 #####
  2740. [u'loss', -0.007816069759428501, 0.033731184899806976, -2.291435956954956, 1.2572479248046875, -3.6175846616970375e-05, 0.3474588096141815, 0.18148478865623474]
  2741. receiving
  2742. ##### Sending to neptune:  active_relus :  0.199473697742 , 8790208.01 #####
  2743. ##### Sending to neptune:  dp_per_s :  0.199473697742 , 89.2401421731 #####
  2744. [u'other', 8790208.01, 89.24014217307855]
  2745. receiving
  2746. [0130 17:41:59 @multigpu.py:323] [p0576]  step: count(507), step_time 1433.45, mean_step_time 1429.27, it/s 0.7
  2747. sending to address tcp://p0112:61216
  2748. ##### Sending to neptune:  online_score :  0.199656606052 , 1.1 #####
  2749. [u'online', 1.1]
  2750. receiving
  2751. [0130 17:42:00 @multigpu.py:323] [p0574]  step: count(501), step_time 1400.88, mean_step_time 1428.05, it/s 0.7
  2752. [0130 17:42:00 @multigpu.py:323] [p0115]  step: count(485), step_time 1468.38, mean_step_time 1433.27, it/s 0.7
  2753. [0130 17:42:00 @multigpu.py:323] [p0576]  step: count(508), step_time 1441.14, mean_step_time 1430.38, it/s 0.7
  2754. [0130 17:42:01 @multigpu.py:323] [p0574]  step: count(502), step_time 1372.72, mean_step_time 1425.65, it/s 0.7
  2755. [0130 17:42:02 @multigpu.py:323] [p0115]  step: count(486), step_time 1390.88, mean_step_time 1430.66, it/s 0.7
  2756. [0130 17:42:02 @multigpu.py:323] [p0576]  step: count(509), step_time 1419.81, mean_step_time 1430.44, it/s 0.7
  2757. [0130 17:42:03 @multigpu.py:323] [p0574]  step: count(503), step_time 1453.52, mean_step_time 1429.07, it/s 0.7
  2758. [0130 17:42:03 @multigpu.py:323] [p0115]  step: count(487), step_time 1401.45, mean_step_time 1426.34, it/s 0.7
  2759. [0130 17:42:03 @multigpu.py:323] [p0576]  step: count(510), step_time 1404.39, mean_step_time 1432.51, it/s 0.7
  2760. [0130 17:42:05 @multigpu.py:323] [p0115]  step: count(488), step_time 1536.76, mean_step_time 1431.83, it/s 0.7
  2761. [0130 17:42:05 @multigpu.py:323] [p0574]  step: count(504), step_time 1569.74, mean_step_time 1431.79, it/s 0.7
  2762. [0130 17:42:05 @multigpu.py:323] [p0576]  step: count(511), step_time 1427.58, mean_step_time 1431.77, it/s 0.7
  2763. sending to address tcp://p0112:61216
  2764. ##### Sending to neptune:  online_score :  0.201240341332 , 1.7 #####
  2765. [u'online', 1.7]
  2766. receiving
  2767. [0130 17:42:06 @multigpu.py:323] [p0574]  step: count(505), step_time 1407.0, mean_step_time 1431.12, it/s 0.7
  2768. [0130 17:42:06 @multigpu.py:323] [p0115]  step: count(489), step_time 1416.23, mean_step_time 1432.34, it/s 0.7
  2769. [0130 17:42:06 @multigpu.py:323] [p0576]  step: count(512), step_time 1428.95, mean_step_time 1431.52, it/s 0.7
  2770. sending to address tcp://p0112:61216
  2771. ##### Sending to neptune:  online_score :  0.201575732165 , 1.6 #####
  2772. [u'online', 1.6]
  2773. receiving
  2774. [0130 17:42:07 @multigpu.py:323] [p0115]  step: count(490), step_time 1366.42, mean_step_time 1427.2, it/s 0.7
  2775. [0130 17:42:07 @multigpu.py:323] [p0576]  step: count(513), step_time 1370.6, mean_step_time 1426.8, it/s 0.7
  2776. [0130 17:42:07 @multigpu.py:323] [p0574]  step: count(506), step_time 1396.16, mean_step_time 1427.82, it/s 0.7
  2777. [0130 17:42:09 @multigpu.py:323] [p0115]  step: count(491), step_time 1405.04, mean_step_time 1426.8, it/s 0.7
  2778. [0130 17:42:09 @multigpu.py:323] [p0574]  step: count(507), step_time 1438.12, mean_step_time 1429.64, it/s 0.7
  2779. [0130 17:42:09 @multigpu.py:323] [p0576]  step: count(514), step_time 1453.14, mean_step_time 1428.03, it/s 0.7
  2780. [0130 17:42:10 @multigpu.py:323] [p0115]  step: count(492), step_time 1436.25, mean_step_time 1428.47, it/s 0.7
  2781. [0130 17:42:10 @multigpu.py:323] [p0576]  step: count(515), step_time 1375.54, mean_step_time 1427.14, it/s 0.7
  2782. [0130 17:42:10 @multigpu.py:323] [p0574]  step: count(508), step_time 1403.88, mean_step_time 1429.95, it/s 0.7
  2783. [0130 17:42:12 @multigpu.py:323] [p0115]  step: count(493), step_time 1390.78, mean_step_time 1426.04, it/s 0.7
  2784. [0130 17:42:12 @multigpu.py:323] [p0576]  step: count(516), step_time 1406.27, mean_step_time 1427.31, it/s 0.7
  2785. [0130 17:42:12 @multigpu.py:323] [p0574]  step: count(509), step_time 1413.02, mean_step_time 1429.63, it/s 0.7
  2786. [0130 17:42:13 @multigpu.py:323] [p0576]  step: count(517), step_time 1391.98, mean_step_time 1427.35, it/s 0.7
  2787. [0130 17:42:13 @multigpu.py:323] [p0115]  step: count(494), step_time 1422.91, mean_step_time 1425.55, it/s 0.7
  2788. [0130 17:42:13 @multigpu.py:323] [p0574]  step: count(510), step_time 1383.5, mean_step_time 1428.23, it/s 0.7
  2789. [0130 17:42:14 @multigpu.py:323] [p0576]  step: count(518), step_time 1410.97, mean_step_time 1428.1, it/s 0.7
  2790. [0130 17:42:14 @multigpu.py:323] [p0574]  step: count(511), step_time 1423.99, mean_step_time 1428.15, it/s 0.7
  2791. [0130 17:42:14 @multigpu.py:323] [p0115]  step: count(495), step_time 1461.77, mean_step_time 1426.17, it/s 0.7
  2792. [0130 17:42:16 @multigpu.py:323] [p0574]  step: count(512), step_time 1455.42, mean_step_time 1431.07, it/s 0.7
  2793. [0130 17:42:16 @multigpu.py:323] [p0576]  step: count(519), step_time 1486.84, mean_step_time 1430.1, it/s 0.7
  2794. [0130 17:42:16 @multigpu.py:323] [p0115]  step: count(496), step_time 1454.06, mean_step_time 1426.13, it/s 0.7
  2795. [0130 17:42:17 @multigpu.py:323] [p0574]  step: count(513), step_time 1404.82, mean_step_time 1431.14, it/s 0.7
  2796. [0130 17:42:17 @multigpu.py:323] [p0576]  step: count(520), step_time 1438.83, mean_step_time 1430.71, it/s 0.7
  2797. [0130 17:42:17 @multigpu.py:323] [p0115]  step: count(497), step_time 1474.65, mean_step_time 1430.33, it/s 0.7
  2798. [0130 17:42:19 @multigpu.py:323] [p0574]  step: count(514), step_time 1431.24, mean_step_time 1431.06, it/s 0.7
  2799. [0130 17:42:19 @multigpu.py:323] [p0576]  step: count(521), step_time 1396.57, mean_step_time 1425.84, it/s 0.7
  2800. [0130 17:42:19 @multigpu.py:323] [p0115]  step: count(498), step_time 1381.51, mean_step_time 1430.91, it/s 0.7
  2801. [0130 17:42:20 @multigpu.py:323] [p0576]  step: count(522), step_time 1417.83, mean_step_time 1423.06, it/s 0.7
  2802. [0130 17:42:20 @multigpu.py:323] [p0574]  step: count(515), step_time 1437.48, mean_step_time 1430.87, it/s 0.7
  2803. [0130 17:42:20 @multigpu.py:323] [p0115]  step: count(499), step_time 1391.41, mean_step_time 1430.45, it/s 0.7
  2804. [0130 17:42:22 @multigpu.py:323] [p0574]  step: count(516), step_time 1395.2, mean_step_time 1429.3, it/s 0.7
  2805. [0130 17:42:22 @multigpu.py:323] [p0576]  step: count(523), step_time 1412.15, mean_step_time 1422.81, it/s 0.7
  2806. [0130 17:42:22 @multigpu.py:323] [p0115]  step: count(500), step_time 1470.12, mean_step_time 1430.72, it/s 0.7
  2807. sending debugging info...
  2808. sending to address tcp://p0112:61216
  2809. ##### Sending to neptune:  mean_delay :  0.205827626652 , 0.0 #####
  2810. sending to address tcp://p0112:61216
  2811. ##### Sending to neptune:  max_delay :  0.205827626652 , -0.0 #####
  2812. ##### Sending to neptune:  min_delay :  0.205827626652 , -0.0 #####
  2813. [u'delays', [0.0, -0.0, -0.0]]
  2814. receiving
  2815. ##### Sending to neptune:  cost :  0.205828117198 , -0.0121910655871 #####
  2816. sending to address tcp://p0112:61216
  2817. ##### Sending to neptune:  policy_loss :  0.205828117198 , -0.40860247612 #####
  2818. ##### Sending to neptune:  xentropy_loss :  0.205828117198 , -2.29158711433 #####
  2819. ##### Sending to neptune:  value_loss :  0.205828117198 , 1.1397330761 #####
  2820. ##### Sending to neptune:  advantage :  0.205828117198 , 0.00180070789065 #####
  2821. ##### Sending to neptune:  pred_reward :  0.205828117198 , 0.358477056026 #####
  2822. ##### Sending to neptune:  max_logit :  0.205828117198 , 0.181056678295 #####
  2823. [u'loss', -0.012191065587103367, -0.4086024761199951, -2.2915871143341064, 1.139733076095581, 0.0018007078906521201, 0.35847705602645874, 0.1810566782951355]
  2824. receiving
  2825. ##### Sending to neptune:  active_relus :  0.205828589996 , 8779679.08 #####
  2826. ##### Sending to neptune:  dp_per_s :  0.205828589996 , 89.1530482324 #####
  2827. [u'other', 8779679.08, 89.15304823242096]
  2828. receiving
  2829. [0130 17:42:23 @multigpu.py:323] [p0574]  step: count(517), step_time 1393.95, mean_step_time 1425.2, it/s 0.7
  2830. [0130 17:42:23 @multigpu.py:323] [p0576]  step: count(524), step_time 1418.22, mean_step_time 1421.53, it/s 0.7
  2831. [0130 17:42:23 @multigpu.py:323] [p0115]  step: count(501), step_time 1392.3, mean_step_time 1427.49, it/s 0.7
  2832. [0130 17:42:24 @multigpu.py:323] [p0574]  step: count(518), step_time 1433.31, mean_step_time 1425.68, it/s 0.7
  2833. [0130 17:42:24 @multigpu.py:323] [p0576]  step: count(525), step_time 1478.57, mean_step_time 1422.96, it/s 0.7
  2834. [0130 17:42:24 @multigpu.py:323] [p0115]  step: count(502), step_time 1453.34, mean_step_time 1428.07, it/s 0.7
  2835. sending to address tcp://p0112:61216
  2836. ##### Sending to neptune:  online_score :  0.206725753281 , 0.9 #####
  2837. [u'online', 0.9]
  2838. receiving
  2839. [0130 17:42:26 @multigpu.py:323] [p0574]  step: count(519), step_time 1437.16, mean_step_time 1423.38, it/s 0.7
  2840. [0130 17:42:26 @multigpu.py:323] [p0576]  step: count(526), step_time 1381.81, mean_step_time 1419.73, it/s 0.7
  2841. [0130 17:42:26 @multigpu.py:323] [p0115]  step: count(503), step_time 1398.28, mean_step_time 1424.65, it/s 0.7
  2842. [0130 17:42:27 @multigpu.py:323] [p0576]  step: count(527), step_time 1381.54, mean_step_time 1417.14, it/s 0.71
  2843. [0130 17:42:27 @multigpu.py:323] [p0574]  step: count(520), step_time 1449.36, mean_step_time 1425.02, it/s 0.7
  2844. [0130 17:42:27 @multigpu.py:323] [p0115]  step: count(504), step_time 1423.18, mean_step_time 1426.79, it/s 0.7
  2845. [0130 17:42:29 @multigpu.py:323] [p0576]  step: count(528), step_time 1415.51, mean_step_time 1415.86, it/s 0.71
  2846. [0130 17:42:29 @multigpu.py:323] [p0115]  step: count(505), step_time 1373.35, mean_step_time 1422.04, it/s 0.7
  2847. [0130 17:42:29 @multigpu.py:323] [p0574]  step: count(521), step_time 1408.41, mean_step_time 1425.4, it/s 0.7
  2848. [0130 17:42:30 @multigpu.py:323] [p0576]  step: count(529), step_time 1428.52, mean_step_time 1416.29, it/s 0.71
  2849. [0130 17:42:30 @multigpu.py:323] [p0574]  step: count(522), step_time 1431.4, mean_step_time 1428.33, it/s 0.7
  2850. sending to address tcp://p0112:61216
  2851. [0130 17:42:30 @multigpu.py:323] [p0115]  step: count(506), step_time 1448.29, mean_step_time 1424.91, it/s 0.7
  2852. ##### Sending to neptune:  online_score :  0.208189289702 , 1.6 #####
  2853. [u'online', 1.6]
  2854. receiving
  2855. [0130 17:42:31 @multigpu.py:323] [p0576]  step: count(530), step_time 1397.03, mean_step_time 1415.92, it/s 0.71
  2856. [0130 17:42:32 @multigpu.py:323] [p0115]  step: count(507), step_time 1437.19, mean_step_time 1426.69, it/s 0.7
  2857. [0130 17:42:32 @multigpu.py:323] [p0574]  step: count(523), step_time 1454.59, mean_step_time 1428.39, it/s 0.7
  2858. [0130 17:42:33 @multigpu.py:323] [p0574]  step: count(524), step_time 1493.74, mean_step_time 1424.59, it/s 0.7
  2859. [0130 17:42:33 @multigpu.py:323] [p0576]  step: count(531), step_time 1604.96, mean_step_time 1424.79, it/s 0.7
  2860. [0130 17:42:33 @multigpu.py:323] [p0115]  step: count(508), step_time 1498.7, mean_step_time 1424.79, it/s 0.7
  2861. sending to address tcp://p0112:61216
  2862. ##### Sending to neptune:  online_score :  0.209143604438 , 1.8 #####
  2863. [u'online', 1.8]
  2864. receiving
  2865. [0130 17:42:34 @multigpu.py:323] [p0574]  step: count(525), step_time 1396.72, mean_step_time 1424.07, it/s 0.7
  2866. [0130 17:42:34 @multigpu.py:323] [p0115]  step: count(509), step_time 1457.94, mean_step_time 1426.88, it/s 0.7
  2867. [0130 17:42:34 @multigpu.py:323] [p0576]  step: count(532), step_time 1484.47, mean_step_time 1427.57, it/s 0.7
  2868. [0130 17:42:36 @multigpu.py:323] [p0574]  step: count(526), step_time 1379.54, mean_step_time 1423.24, it/s 0.7
  2869. [0130 17:42:36 @multigpu.py:323] [p0115]  step: count(510), step_time 1384.93, mean_step_time 1427.8, it/s 0.7
  2870. [0130 17:42:36 @multigpu.py:323] [p0576]  step: count(533), step_time 1391.0, mean_step_time 1428.59, it/s 0.7
  2871. [0130 17:42:37 @multigpu.py:323] [p0574]  step: count(527), step_time 1455.82, mean_step_time 1424.13, it/s 0.7
  2872. [0130 17:42:37 @multigpu.py:323] [p0576]  step: count(534), step_time 1399.56, mean_step_time 1425.91, it/s 0.7
  2873. [0130 17:42:37 @multigpu.py:323] [p0115]  step: count(511), step_time 1458.11, mean_step_time 1430.45, it/s 0.7
  2874. [0130 17:42:39 @multigpu.py:323] [p0115]  step: count(512), step_time 1391.15, mean_step_time 1428.2, it/s 0.7
  2875. [0130 17:42:39 @multigpu.py:323] [p0574]  step: count(528), step_time 1511.41, mean_step_time 1429.5, it/s 0.7
  2876. [0130 17:42:39 @multigpu.py:323] [p0576]  step: count(535), step_time 1479.66, mean_step_time 1431.11, it/s 0.7
  2877. [0130 17:42:40 @multigpu.py:323] [p0576]  step: count(536), step_time 1398.88, mean_step_time 1430.74, it/s 0.7
  2878. [0130 17:42:40 @multigpu.py:323] [p0115]  step: count(513), step_time 1464.78, mean_step_time 1431.9, it/s 0.7
  2879. [0130 17:42:40 @multigpu.py:323] [p0574]  step: count(529), step_time 1445.78, mean_step_time 1431.14, it/s 0.7
  2880. [0130 17:42:42 @multigpu.py:323] [p0115]  step: count(514), step_time 1391.88, mean_step_time 1430.35, it/s 0.7
  2881. [0130 17:42:42 @multigpu.py:323] [p0574]  step: count(530), step_time 1430.96, mean_step_time 1433.52, it/s 0.7
  2882. [0130 17:42:42 @multigpu.py:323] [p0576]  step: count(537), step_time 1485.03, mean_step_time 1435.4, it/s 0.7
  2883. [0130 17:42:43 @multigpu.py:323] [p0115]  step: count(515), step_time 1505.01, mean_step_time 1432.51, it/s 0.7
  2884. [0130 17:42:43 @multigpu.py:323] [p0574]  step: count(531), step_time 1441.49, mean_step_time 1434.39, it/s 0.7
  2885. [0130 17:42:43 @multigpu.py:323] [p0576]  step: count(538), step_time 1469.73, mean_step_time 1438.34, it/s 0.7
  2886. [0130 17:42:44 @multigpu.py:323] [p0115]  step: count(516), step_time 1386.08, mean_step_time 1429.11, it/s 0.7
  2887. [0130 17:42:44 @multigpu.py:323] [p0574]  step: count(532), step_time 1381.01, mean_step_time 1430.67, it/s 0.7
  2888. [0130 17:42:44 @multigpu.py:323] [p0576]  step: count(539), step_time 1360.79, mean_step_time 1432.03, it/s 0.7
  2889. [0130 17:42:46 @multigpu.py:323] [p0574]  step: count(533), step_time 1399.77, mean_step_time 1430.42, it/s 0.7
  2890. [0130 17:42:46 @multigpu.py:323] [p0576]  step: count(540), step_time 1413.46, mean_step_time 1430.76, it/s 0.7
  2891. [0130 17:42:46 @multigpu.py:323] [p0115]  step: count(517), step_time 1497.86, mean_step_time 1430.27, it/s 0.7
  2892. [0130 17:42:47 @multigpu.py:323] [p0574]  step: count(534), step_time 1418.43, mean_step_time 1429.78, it/s 0.7
  2893. [0130 17:42:47 @multigpu.py:323] [p0576]  step: count(541), step_time 1464.92, mean_step_time 1434.18, it/s 0.7
  2894. [0130 17:42:47 @multigpu.py:323] [p0115]  step: count(518), step_time 1473.97, mean_step_time 1434.89, it/s 0.7
  2895. [0130 17:42:49 @multigpu.py:323] [p0574]  step: count(535), step_time 1418.27, mean_step_time 1428.82, it/s 0.7
  2896. [0130 17:42:49 @multigpu.py:323] [p0576]  step: count(542), step_time 1427.37, mean_step_time 1434.66, it/s 0.7
  2897. [0130 17:42:49 @multigpu.py:323] [p0115]  step: count(519), step_time 1454.96, mean_step_time 1438.07, it/s 0.7
  2898. [0130 17:42:50 @multigpu.py:323] [p0574]  step: count(536), step_time 1423.84, mean_step_time 1430.25, it/s 0.7
  2899. [0130 17:42:50 @multigpu.py:323] [p0576]  step: count(543), step_time 1444.12, mean_step_time 1436.26, it/s 0.7
  2900. [0130 17:42:50 @multigpu.py:323] [p0115]  step: count(520), step_time 1469.49, mean_step_time 1438.04, it/s 0.7
  2901. sending to address tcp://p0112:61216
  2902. ##### Sending to neptune:  online_score :  0.213829125497 , 1.3 #####
  2903. [u'online', 1.3]
  2904. ##### Sending to neptune:  active_workers :  0.213829183843 , 3 #####
  2905. receiving
  2906. [0130 17:42:51 @multigpu.py:323] [p0574]  step: count(537), step_time 1378.7, mean_step_time 1429.49, it/s 0.7
  2907. [0130 17:42:52 @multigpu.py:323] [p0576]  step: count(544), step_time 1363.11, mean_step_time 1433.5, it/s 0.7
  2908. [0130 17:42:52 @multigpu.py:323] [p0115]  step: count(521), step_time 1388.45, mean_step_time 1437.85, it/s 0.7
  2909. [0130 17:42:53 @multigpu.py:323] [p0574]  step: count(538), step_time 1451.19, mean_step_time 1430.38, it/s 0.7
  2910. [0130 17:42:53 @multigpu.py:323] [p0576]  step: count(545), step_time 1422.98, mean_step_time 1430.72, it/s 0.7
  2911. [0130 17:42:53 @multigpu.py:323] [p0115]  step: count(522), step_time 1421.59, mean_step_time 1436.26, it/s 0.7
  2912. [0130 17:42:54 @multigpu.py:323] [p0574]  step: count(539), step_time 1439.49, mean_step_time 1430.5, it/s 0.7
  2913. [0130 17:42:54 @multigpu.py:323] [p0576]  step: count(546), step_time 1419.38, mean_step_time 1432.6, it/s 0.7
  2914. [0130 17:42:55 @multigpu.py:323] [p0115]  step: count(523), step_time 1437.96, mean_step_time 1438.24, it/s 0.7
  2915. [0130 17:42:56 @multigpu.py:323] [p0574]  step: count(540), step_time 1451.98, mean_step_time 1430.63, it/s 0.7
  2916. [0130 17:42:56 @multigpu.py:323] [p0576]  step: count(547), step_time 1419.85, mean_step_time 1434.52, it/s 0.7
  2917. [0130 17:42:56 @multigpu.py:323] [p0115]  step: count(524), step_time 1419.43, mean_step_time 1438.06, it/s 0.7
  2918. sending to address tcp://p0112:61216
  2919. ##### Sending to neptune:  online_score :  0.21545850469 , 2.3 #####
  2920. [u'online', 2.3]
  2921. receiving
  2922. [0130 17:42:57 @multigpu.py:323] [p0574]  step: count(541), step_time 1421.9, mean_step_time 1431.3, it/s 0.7
  2923. [0130 17:42:57 @multigpu.py:323] [p0576]  step: count(548), step_time 1434.48, mean_step_time 1435.46, it/s 0.7
  2924. [0130 17:42:57 @multigpu.py:323] [p0115]  step: count(525), step_time 1418.32, mean_step_time 1440.3, it/s 0.69
  2925. [0130 17:42:59 @multigpu.py:323] [p0574]  step: count(542), step_time 1432.74, mean_step_time 1431.37, it/s 0.7
  2926. [0130 17:42:59 @multigpu.py:323] [p0576]  step: count(549), step_time 1452.17, mean_step_time 1436.65, it/s 0.7
  2927. [0130 17:42:59 @multigpu.py:323] [p0115]  step: count(526), step_time 1431.79, mean_step_time 1439.48, it/s 0.69
  2928. [0130 17:43:00 @multigpu.py:323] [p0574]  step: count(543), step_time 1390.72, mean_step_time 1428.18, it/s 0.7
  2929. [0130 17:43:00 @multigpu.py:323] [p0576]  step: count(550), step_time 1431.72, mean_step_time 1438.38, it/s 0.7
  2930. [0130 17:43:00 @multigpu.py:323] [p0115]  step: count(527), step_time 1469.17, mean_step_time 1441.08, it/s 0.69
  2931. [0130 17:43:02 @multigpu.py:323] [p0115]  step: count(528), step_time 1439.59, mean_step_time 1438.12, it/s 0.7
  2932. [0130 17:43:02 @multigpu.py:323] [p0576]  step: count(551), step_time 1598.78, mean_step_time 1438.07, it/s 0.7
  2933. [0130 17:43:02 @multigpu.py:323] [p0574]  step: count(544), step_time 1693.96, mean_step_time 1438.19, it/s 0.7
  2934. [0130 17:43:03 @multigpu.py:323] [p0576]  step: count(552), step_time 1420.36, mean_step_time 1434.87, it/s 0.7
  2935. [0130 17:43:03 @multigpu.py:323] [p0574]  step: count(545), step_time 1434.88, mean_step_time 1440.09, it/s 0.69
  2936. [0130 17:43:03 @multigpu.py:323] [p0115]  step: count(529), step_time 1502.04, mean_step_time 1440.33, it/s 0.69
  2937. [0130 17:43:05 @multigpu.py:323] [p0576]  step: count(553), step_time 1381.06, mean_step_time 1434.37, it/s 0.7
  2938. [0130 17:43:05 @multigpu.py:323] [p0574]  step: count(546), step_time 1432.99, mean_step_time 1442.77, it/s 0.69
  2939. [0130 17:43:05 @multigpu.py:323] [p0115]  step: count(530), step_time 1425.52, mean_step_time 1442.36, it/s 0.69
  2940. sending to address tcp://p0112:61216
  2941. ##### Sending to neptune:  online_score :  0.217906641629 , 1.8 #####
  2942. [u'online', 1.8]
  2943. receiving
  2944. [0130 17:43:06 @multigpu.py:323] [p0576]  step: count(554), step_time 1442.1, mean_step_time 1436.5, it/s 0.7
  2945. [0130 17:43:06 @multigpu.py:323] [p0574]  step: count(547), step_time 1473.52, mean_step_time 1443.65, it/s 0.69
  2946. [0130 17:43:06 @multigpu.py:323] [p0115]  step: count(531), step_time 1427.41, mean_step_time 1440.82, it/s 0.69
  2947. [0130 17:43:07 @multigpu.py:323] [p0576]  step: count(555), step_time 1424.39, mean_step_time 1433.73, it/s 0.7
  2948. [0130 17:43:08 @multigpu.py:323] [p0574]  step: count(548), step_time 1439.58, mean_step_time 1440.06, it/s 0.69
  2949. [0130 17:43:08 @multigpu.py:323] [p0115]  step: count(532), step_time 1430.38, mean_step_time 1442.78, it/s 0.69
  2950. [0130 17:43:09 @multigpu.py:323] [p0576]  step: count(556), step_time 1518.46, mean_step_time 1439.71, it/s 0.69
  2951. [0130 17:43:09 @multigpu.py:323] [p0574]  step: count(549), step_time 1427.17, mean_step_time 1439.13, it/s 0.69
  2952. [0130 17:43:09 @multigpu.py:323] [p0115]  step: count(533), step_time 1443.05, mean_step_time 1441.7, it/s 0.69
  2953. [0130 17:43:10 @multigpu.py:323] [p0576]  step: count(557), step_time 1374.29, mean_step_time 1434.18, it/s 0.7
  2954. [0130 17:43:10 @multigpu.py:323] [p0115]  step: count(534), step_time 1428.15, mean_step_time 1443.51, it/s 0.69
  2955. [0130 17:43:10 @multigpu.py:323] [p0574]  step: count(550), step_time 1473.98, mean_step_time 1441.28, it/s 0.69
  2956. [0130 17:43:12 @multigpu.py:323] [p0576]  step: count(558), step_time 1458.34, mean_step_time 1433.61, it/s 0.7
  2957. [0130 17:43:12 @multigpu.py:323] [p0574]  step: count(551), step_time 1400.64, mean_step_time 1439.24, it/s 0.69
  2958. [0130 17:43:12 @multigpu.py:323] [p0115]  step: count(535), step_time 1427.15, mean_step_time 1439.62, it/s 0.69
  2959. [0130 17:43:13 @multigpu.py:323] [p0576]  step: count(559), step_time 1451.05, mean_step_time 1438.12, it/s 0.7
  2960. [0130 17:43:13 @multigpu.py:323] [p0115]  step: count(536), step_time 1428.3, mean_step_time 1441.73, it/s 0.69
  2961. [0130 17:43:13 @multigpu.py:323] [p0574]  step: count(552), step_time 1464.41, mean_step_time 1443.41, it/s 0.69
  2962. [0130 17:43:15 @multigpu.py:323] [p0576]  step: count(560), step_time 1439.48, mean_step_time 1439.42, it/s 0.69
  2963. [0130 17:43:15 @multigpu.py:323] [p0115]  step: count(537), step_time 1441.73, mean_step_time 1438.92, it/s 0.69
  2964. [0130 17:43:15 @multigpu.py:323] [p0574]  step: count(553), step_time 1434.15, mean_step_time 1445.13, it/s 0.69
  2965. [0130 17:43:16 @multigpu.py:323] [p0576]  step: count(561), step_time 1380.15, mean_step_time 1435.18, it/s 0.7
  2966. [0130 17:43:16 @multigpu.py:323] [p0574]  step: count(554), step_time 1402.8, mean_step_time 1444.35, it/s 0.69
  2967. [0130 17:43:16 @multigpu.py:323] [p0115]  step: count(538), step_time 1504.04, mean_step_time 1440.43, it/s 0.69
  2968. [0130 17:43:17 @multigpu.py:323] [p0576]  step: count(562), step_time 1414.47, mean_step_time 1434.54, it/s 0.7
  2969. [0130 17:43:18 @multigpu.py:323] [p0574]  step: count(555), step_time 1400.29, mean_step_time 1443.45, it/s 0.69
  2970. [0130 17:43:18 @multigpu.py:323] [p0115]  step: count(539), step_time 1493.64, mean_step_time 1442.36, it/s 0.69
  2971. [0130 17:43:19 @multigpu.py:323] [p0576]  step: count(563), step_time 1459.15, mean_step_time 1435.29, it/s 0.7
  2972. [0130 17:43:19 @multigpu.py:323] [p0574]  step: count(556), step_time 1398.45, mean_step_time 1442.18, it/s 0.69
  2973. [0130 17:43:19 @multigpu.py:323] [p0115]  step: count(540), step_time 1440.72, mean_step_time 1440.92, it/s 0.69
  2974. sending to address tcp://p0112:61216
  2975. ##### Sending to neptune:  online_score :  0.221927292479 , 1.7 #####
  2976. [u'online', 1.7]
  2977. receiving
  2978. [0130 17:43:20 @multigpu.py:323] [p0574]  step: count(557), step_time 1431.14, mean_step_time 1444.8, it/s 0.69
  2979. [0130 17:43:20 @multigpu.py:323] [p0576]  step: count(564), step_time 1454.93, mean_step_time 1439.88, it/s 0.69
  2980. [0130 17:43:21 @multigpu.py:323] [p0115]  step: count(541), step_time 1474.27, mean_step_time 1445.21, it/s 0.69
  2981. [0130 17:43:22 @multigpu.py:323] [p0574]  step: count(558), step_time 1400.69, mean_step_time 1442.27, it/s 0.69
  2982. [0130 17:43:22 @multigpu.py:323] [p0576]  step: count(565), step_time 1448.58, mean_step_time 1441.16, it/s 0.69
  2983. [0130 17:43:22 @multigpu.py:323] [p0115]  step: count(542), step_time 1401.76, mean_step_time 1444.22, it/s 0.69
  2984. sending to address tcp://p0112:61216
  2985. ##### Sending to neptune:  online_score :  0.222741466893 , 1.1 #####
  2986. [u'online', 1.1]
  2987. receiving
  2988. [0130 17:43:23 @multigpu.py:323] [p0576]  step: count(566), step_time 1423.63, mean_step_time 1441.37, it/s 0.69
  2989. [0130 17:43:23 @multigpu.py:323] [p0574]  step: count(559), step_time 1486.25, mean_step_time 1444.61, it/s 0.69
  2990. [0130 17:43:23 @multigpu.py:323] [p0115]  step: count(543), step_time 1438.65, mean_step_time 1444.26, it/s 0.69
  2991. [0130 17:43:25 @multigpu.py:323] [p0576]  step: count(567), step_time 1395.37, mean_step_time 1440.15, it/s 0.69
  2992. [0130 17:43:25 @multigpu.py:323] [p0574]  step: count(560), step_time 1489.61, mean_step_time 1446.49, it/s 0.69
  2993. [0130 17:43:25 @multigpu.py:323] [p0115]  step: count(544), step_time 1469.23, mean_step_time 1446.75, it/s 0.69
  2994. [0130 17:43:26 @multigpu.py:323] [p0576]  step: count(568), step_time 1395.78, mean_step_time 1438.21, it/s 0.7
  2995. [0130 17:43:26 @multigpu.py:323] [p0574]  step: count(561), step_time 1387.96, mean_step_time 1444.8, it/s 0.69
  2996. [0130 17:43:26 @multigpu.py:323] [p0115]  step: count(545), step_time 1483.14, mean_step_time 1449.99, it/s 0.69
  2997. [0130 17:43:28 @multigpu.py:323] [p0576]  step: count(569), step_time 1442.61, mean_step_time 1437.74, it/s 0.7
  2998. [0130 17:43:28 @multigpu.py:323] [p0574]  step: count(562), step_time 1438.72, mean_step_time 1445.1, it/s 0.69
  2999. [0130 17:43:28 @multigpu.py:323] [p0115]  step: count(546), step_time 1477.18, mean_step_time 1452.26, it/s 0.69
  3000. sending to address tcp://p0112:61216
  3001. ##### Sending to neptune:  online_score :  0.224450801346 , 1.6 #####
  3002. [u'online', 1.6]
  3003. receiving
  3004. [0130 17:43:29 @multigpu.py:323] [p0576]  step: count(570), step_time 1441.42, mean_step_time 1438.22, it/s 0.7
  3005. [0130 17:43:29 @multigpu.py:323] [p0574]  step: count(563), step_time 1451.3, mean_step_time 1448.12, it/s 0.69
  3006. [0130 17:43:29 @multigpu.py:323] [p0115]  step: count(547), step_time 1422.33, mean_step_time 1449.91, it/s 0.69
  3007. [0130 17:43:31 @multigpu.py:323] [p0115]  step: count(548), step_time 1458.19, mean_step_time 1450.84, it/s 0.69
  3008. [0130 17:43:31 @multigpu.py:323] [p0574]  step: count(564), step_time 1747.66, mean_step_time 1450.81, it/s 0.69
  3009. [0130 17:43:31 @multigpu.py:323] [p0576]  step: count(571), step_time 1850.98, mean_step_time 1450.83, it/s 0.69
  3010. [0130 17:43:32 @multigpu.py:323] [p0576]  step: count(572), step_time 1377.4, mean_step_time 1448.68, it/s 0.69
  3011. [0130 17:43:32 @multigpu.py:323] [p0115]  step: count(549), step_time 1430.69, mean_step_time 1447.28, it/s 0.69
  3012. [0130 17:43:32 @multigpu.py:323] [p0574]  step: count(565), step_time 1487.1, mean_step_time 1453.42, it/s 0.69
  3013. [0130 17:43:34 @multigpu.py:323] [p0576]  step: count(573), step_time 1371.86, mean_step_time 1448.22, it/s 0.69
  3014. [0130 17:43:34 @multigpu.py:323] [p0115]  step: count(550), step_time 1427.34, mean_step_time 1447.37, it/s 0.69
  3015. [0130 17:43:34 @multigpu.py:323] [p0574]  step: count(566), step_time 1374.94, mean_step_time 1450.52, it/s 0.69
  3016. [0130 17:43:35 @multigpu.py:323] [p0576]  step: count(574), step_time 1433.33, mean_step_time 1447.78, it/s 0.69
  3017. [0130 17:43:35 @multigpu.py:323] [p0115]  step: count(551), step_time 1411.06, mean_step_time 1446.55, it/s 0.69
  3018. [0130 17:43:35 @multigpu.py:323] [p0574]  step: count(567), step_time 1443.21, mean_step_time 1449.0, it/s 0.69
  3019. [0130 17:43:36 @multigpu.py:323] [p0576]  step: count(575), step_time 1403.55, mean_step_time 1446.74, it/s 0.69
  3020. [0130 17:43:37 @multigpu.py:323] [p0115]  step: count(552), step_time 1421.4, mean_step_time 1446.1, it/s 0.69
  3021. [0130 17:43:37 @multigpu.py:323] [p0574]  step: count(568), step_time 1440.96, mean_step_time 1449.07, it/s 0.69
  3022. [0130 17:43:38 @multigpu.py:323] [p0576]  step: count(576), step_time 1403.86, mean_step_time 1441.01, it/s 0.69
  3023. [0130 17:43:38 @multigpu.py:323] [p0115]  step: count(553), step_time 1453.49, mean_step_time 1446.62, it/s 0.69
  3024. [0130 17:43:38 @multigpu.py:323] [p0574]  step: count(569), step_time 1424.56, mean_step_time 1448.94, it/s 0.69
  3025. [0130 17:43:39 @multigpu.py:323] [p0576]  step: count(577), step_time 1452.25, mean_step_time 1444.91, it/s 0.69
  3026. [0130 17:43:39 @multigpu.py:323] [p0115]  step: count(554), step_time 1380.55, mean_step_time 1444.24, it/s 0.69
  3027. [0130 17:43:39 @multigpu.py:323] [p0574]  step: count(570), step_time 1430.86, mean_step_time 1446.78, it/s 0.69
  3028. [0130 17:43:41 @multigpu.py:323] [p0576]  step: count(578), step_time 1408.74, mean_step_time 1442.43, it/s 0.69
  3029. [0130 17:43:41 @multigpu.py:323] [p0115]  step: count(555), step_time 1435.03, mean_step_time 1444.64, it/s 0.69
  3030. [0130 17:43:41 @multigpu.py:323] [p0574]  step: count(571), step_time 1406.29, mean_step_time 1447.07, it/s 0.69
  3031. [0130 17:43:42 @multigpu.py:323] [p0576]  step: count(579), step_time 1439.34, mean_step_time 1441.84, it/s 0.69
  3032. [0130 17:43:42 @multigpu.py:323] [p0115]  step: count(556), step_time 1421.63, mean_step_time 1444.3, it/s 0.69
  3033. [0130 17:43:42 @multigpu.py:323] [p0574]  step: count(572), step_time 1442.9, mean_step_time 1445.99, it/s 0.69
  3034. [0130 17:43:43 @multigpu.py:323] [p0576]  step: count(580), step_time 1380.83, mean_step_time 1438.91, it/s 0.69
  3035. [0130 17:43:44 @multigpu.py:323] [p0115]  step: count(557), step_time 1395.92, mean_step_time 1442.01, it/s 0.69
  3036. [0130 17:43:44 @multigpu.py:323] [p0574]  step: count(573), step_time 1402.75, mean_step_time 1444.42, it/s 0.69
  3037. [0130 17:43:45 @multigpu.py:323] [p0576]  step: count(581), step_time 1417.78, mean_step_time 1440.79, it/s 0.69
  3038. [0130 17:43:45 @multigpu.py:323] [p0574]  step: count(574), step_time 1430.28, mean_step_time 1445.8, it/s 0.69
  3039. [0130 17:43:45 @multigpu.py:323] [p0115]  step: count(558), step_time 1527.13, mean_step_time 1443.17, it/s 0.69
  3040. [0130 17:43:46 @multigpu.py:323] [p0576]  step: count(582), step_time 1476.98, mean_step_time 1443.92, it/s 0.69
  3041. [0130 17:43:47 @multigpu.py:323] [p0115]  step: count(559), step_time 1452.16, mean_step_time 1441.09, it/s 0.69
  3042. [0130 17:43:47 @multigpu.py:323] [p0574]  step: count(575), step_time 1483.65, mean_step_time 1449.96, it/s 0.69
  3043. [0130 17:43:48 @multigpu.py:323] [p0576]  step: count(583), step_time 1433.33, mean_step_time 1442.63, it/s 0.69
  3044. [0130 17:43:48 @multigpu.py:323] [p0115]  step: count(560), step_time 1500.53, mean_step_time 1444.08, it/s 0.69
  3045. [0130 17:43:48 @multigpu.py:323] [p0574]  step: count(576), step_time 1490.74, mean_step_time 1454.58, it/s 0.69
  3046. [0130 17:43:49 @multigpu.py:323] [p0576]  step: count(584), step_time 1371.12, mean_step_time 1438.44, it/s 0.7
  3047. [0130 17:43:50 @multigpu.py:323] [p0115]  step: count(561), step_time 1460.94, mean_step_time 1443.42, it/s 0.69
  3048. [0130 17:43:50 @multigpu.py:323] [p0574]  step: count(577), step_time 1482.66, mean_step_time 1457.15, it/s 0.69
  3049. [0130 17:43:51 @multigpu.py:323] [p0576]  step: count(585), step_time 1448.1, mean_step_time 1438.41, it/s 0.7
  3050. [0130 17:43:51 @multigpu.py:323] [p0115]  step: count(562), step_time 1387.06, mean_step_time 1442.68, it/s 0.69
  3051. [0130 17:43:51 @multigpu.py:323] [p0574]  step: count(578), step_time 1464.13, mean_step_time 1460.33, it/s 0.68
  3052. sending to address tcp://p0112:61216
  3053. ##### Sending to neptune:  online_score :  0.230837176641 , 1.8 #####
  3054. [u'online', 1.8]
  3055. ##### Sending to neptune:  active_workers :  0.230837286645 , 3 #####
  3056. receiving
  3057. [0130 17:43:52 @multigpu.py:323] [p0576]  step: count(586), step_time 1444.38, mean_step_time 1439.45, it/s 0.69
  3058. [0130 17:43:52 @multigpu.py:323] [p0115]  step: count(563), step_time 1462.85, mean_step_time 1443.89, it/s 0.69
  3059. [0130 17:43:52 @multigpu.py:323] [p0574]  step: count(579), step_time 1430.18, mean_step_time 1457.52, it/s 0.69
  3060. [0130 17:43:54 @multigpu.py:323] [p0576]  step: count(587), step_time 1440.66, mean_step_time 1441.71, it/s 0.69
  3061. [0130 17:43:54 @multigpu.py:323] [p0115]  step: count(564), step_time 1395.56, mean_step_time 1440.21, it/s 0.69
  3062. [0130 17:43:54 @multigpu.py:323] [p0574]  step: count(580), step_time 1406.88, mean_step_time 1453.39, it/s 0.69
  3063. sending to address tcp://p0112:61216
  3064. ##### Sending to neptune:  online_score :  0.23157861299 , 1.7 #####
  3065. [u'online', 1.7]
  3066. receiving
  3067. [0130 17:43:55 @multigpu.py:323] [p0576]  step: count(588), step_time 1370.65, mean_step_time 1440.46, it/s 0.69
  3068. [0130 17:43:55 @multigpu.py:323] [p0115]  step: count(565), step_time 1395.37, mean_step_time 1435.82, it/s 0.7
  3069. [0130 17:43:55 @multigpu.py:323] [p0574]  step: count(581), step_time 1436.96, mean_step_time 1455.84, it/s 0.69
  3070. [0130 17:43:56 @multigpu.py:323] [p0576]  step: count(589), step_time 1421.98, mean_step_time 1439.43, it/s 0.69
  3071. [0130 17:43:57 @multigpu.py:323] [p0115]  step: count(566), step_time 1388.97, mean_step_time 1431.41, it/s 0.7
  3072. [0130 17:43:57 @multigpu.py:323] [p0574]  step: count(582), step_time 1424.21, mean_step_time 1455.11, it/s 0.69
  3073. [0130 17:43:58 @multigpu.py:323] [p0576]  step: count(590), step_time 1406.28, mean_step_time 1437.67, it/s 0.7
  3074. [0130 17:43:58 @multigpu.py:323] [p0115]  step: count(567), step_time 1483.43, mean_step_time 1434.47, it/s 0.7
  3075. [0130 17:43:58 @multigpu.py:323] [p0574]  step: count(583), step_time 1436.96, mean_step_time 1454.39, it/s 0.69
  3076. [0130 17:44:00 @multigpu.py:323] [p0115]  step: count(568), step_time 1544.23, mean_step_time 1438.77, it/s 0.7
  3077. [0130 17:44:00 @multigpu.py:323] [p0574]  step: count(584), step_time 1433.22, mean_step_time 1438.67, it/s 0.7
  3078. [0130 17:44:00 @multigpu.py:323] [p0576]  step: count(591), step_time 1869.94, mean_step_time 1438.62, it/s 0.7
  3079. [0130 17:44:01 @multigpu.py:323] [p0115]  step: count(569), step_time 1396.0, mean_step_time 1437.03, it/s 0.7
  3080. [0130 17:44:01 @multigpu.py:323] [p0576]  step: count(592), step_time 1410.35, mean_step_time 1440.27, it/s 0.69
  3081. [0130 17:44:01 @multigpu.py:323] [p0574]  step: count(585), step_time 1446.87, mean_step_time 1436.66, it/s 0.7
  3082. sending to address tcp://p0112:61216
  3083. ##### Sending to neptune:  online_score :  0.233688505226 , 1.5 #####
  3084. [u'online', 1.5]
  3085. receiving
  3086. [0130 17:44:02 @multigpu.py:323] [p0115]  step: count(570), step_time 1421.33, mean_step_time 1436.73, it/s 0.7
  3087. [0130 17:44:02 @multigpu.py:323] [p0576]  step: count(593), step_time 1417.78, mean_step_time 1442.56, it/s 0.69
  3088. [0130 17:44:02 @multigpu.py:323] [p0574]  step: count(586), step_time 1389.72, mean_step_time 1437.4, it/s 0.7
  3089. [0130 17:44:04 @multigpu.py:323] [p0115]  step: count(571), step_time 1425.36, mean_step_time 1437.45, it/s 0.7
  3090. [0130 17:44:04 @multigpu.py:323] [p0574]  step: count(587), step_time 1438.96, mean_step_time 1437.19, it/s 0.7
  3091. [0130 17:44:04 @multigpu.py:323] [p0576]  step: count(594), step_time 1469.7, mean_step_time 1444.38, it/s 0.69
  3092. [0130 17:44:05 @multigpu.py:323] [p0115]  step: count(572), step_time 1433.1, mean_step_time 1438.03, it/s 0.7
  3093. [0130 17:44:05 @multigpu.py:323] [p0576]  step: count(595), step_time 1457.12, mean_step_time 1447.06, it/s 0.69
  3094. [0130 17:44:05 @multigpu.py:323] [p0574]  step: count(588), step_time 1488.52, mean_step_time 1439.57, it/s 0.69
  3095. [0130 17:44:07 @multigpu.py:323] [p0115]  step: count(573), step_time 1404.14, mean_step_time 1435.56, it/s 0.7
  3096. [0130 17:44:07 @multigpu.py:323] [p0576]  step: count(596), step_time 1434.74, mean_step_time 1448.6, it/s 0.69
  3097. [0130 17:44:07 @multigpu.py:323] [p0574]  step: count(589), step_time 1433.57, mean_step_time 1440.02, it/s 0.69
  3098. [0130 17:44:08 @multigpu.py:323] [p0115]  step: count(574), step_time 1462.73, mean_step_time 1439.67, it/s 0.69
  3099. [0130 17:44:08 @multigpu.py:323] [p0576]  step: count(597), step_time 1438.55, mean_step_time 1447.92, it/s 0.69
  3100. [0130 17:44:08 @multigpu.py:323] [p0574]  step: count(590), step_time 1460.88, mean_step_time 1441.52, it/s 0.69
  3101. [0130 17:44:10 @multigpu.py:323] [p0115]  step: count(575), step_time 1428.03, mean_step_time 1439.32, it/s 0.69
  3102. [0130 17:44:10 @multigpu.py:323] [p0576]  step: count(598), step_time 1411.28, mean_step_time 1448.04, it/s 0.69
  3103. [0130 17:44:10 @multigpu.py:323] [p0574]  step: count(591), step_time 1442.06, mean_step_time 1443.3, it/s 0.69
  3104. [0130 17:44:11 @multigpu.py:323] [p0115]  step: count(576), step_time 1430.0, mean_step_time 1439.74, it/s 0.69
  3105. [0130 17:44:11 @multigpu.py:323] [p0576]  step: count(599), step_time 1431.67, mean_step_time 1447.66, it/s 0.69
  3106. [0130 17:44:11 @multigpu.py:323] [p0574]  step: count(592), step_time 1423.58, mean_step_time 1442.34, it/s 0.69
  3107. [0130 17:44:12 @multigpu.py:323] [p0115]  step: count(577), step_time 1407.09, mean_step_time 1440.3, it/s 0.69
  3108. [0130 17:44:13 @multigpu.py:323] [p0574]  step: count(593), step_time 1393.41, mean_step_time 1441.87, it/s 0.69
  3109. [0130 17:44:13 @multigpu.py:323] [p0576]  step: count(600), step_time 1467.88, mean_step_time 1452.01, it/s 0.69
  3110. sending debugging info...
  3111. sending to address tcp://p0112:61216
  3112. ##### Sending to neptune:  mean_delay :  0.236649152173 , 0.0 #####
  3113. sending to address tcp://p0112:61216
  3114. ##### Sending to neptune:  max_delay :  0.236649152173 , -0.0 #####
  3115. ##### Sending to neptune:  min_delay :  0.236649152173 , -0.0 #####
  3116. [u'delays', [0.0, -0.0, -0.0]]
  3117. receiving
  3118. ##### Sending to neptune:  cost :  0.236649644706 , -0.0012414510129 #####
  3119. sending to address tcp://p0112:61216
  3120. ##### Sending to neptune:  policy_loss :  0.236649644706 , 0.698001503944 #####
  3121. ##### Sending to neptune:  xentropy_loss :  0.236649644706 , -2.29192209244 #####
  3122. ##### Sending to neptune:  value_loss :  0.236649644706 , 1.43501460552 #####
  3123. ##### Sending to neptune:  advantage :  0.236649644706 , -0.00304705230519 #####
  3124. ##### Sending to neptune:  pred_reward :  0.236649644706 , 0.39679774642 #####
  3125. ##### Sending to neptune:  max_logit :  0.236649644706 , 0.181564688683 #####
  3126. [u'loss', -0.0012414510129019618, 0.698001503944397, -2.291922092437744, 1.4350146055221558, -0.0030470523051917553, 0.3967977464199066, 0.18156468868255615]
  3127. receiving
  3128. ##### Sending to neptune:  active_relus :  0.236650113596 , 8806054.65 #####
  3129. ##### Sending to neptune:  dp_per_s :  0.236650113596 , 89.2224737498 #####
  3130. [u'other', 8806054.65, 89.22247374981353]
  3131. receiving
  3132. [0130 17:44:14 @multigpu.py:323] [p0115]  step: count(578), step_time 1487.14, mean_step_time 1438.3, it/s 0.7
  3133. [0130 17:44:14 @multigpu.py:323] [p0574]  step: count(594), step_time 1423.42, mean_step_time 1441.53, it/s 0.69
  3134. [0130 17:44:14 @multigpu.py:323] [p0576]  step: count(601), step_time 1441.36, mean_step_time 1453.19, it/s 0.69
  3135. [0130 17:44:15 @multigpu.py:323] [p0115]  step: count(579), step_time 1395.8, mean_step_time 1435.48, it/s 0.7
  3136. [0130 17:44:15 @multigpu.py:323] [p0574]  step: count(595), step_time 1435.73, mean_step_time 1439.13, it/s 0.69
  3137. [0130 17:44:15 @multigpu.py:323] [p0576]  step: count(602), step_time 1481.95, mean_step_time 1453.44, it/s 0.69
  3138. sending to address tcp://p0112:61216
  3139. ##### Sending to neptune:  online_score :  0.237544338836 , 2.1 #####
  3140. [u'online', 2.1]
  3141. receiving
  3142. [0130 17:44:17 @multigpu.py:323] [p0115]  step: count(580), step_time 1488.18, mean_step_time 1434.87, it/s 0.7
  3143. [0130 17:44:17 @multigpu.py:323] [p0574]  step: count(596), step_time 1435.03, mean_step_time 1436.35, it/s 0.7
  3144. [0130 17:44:17 @multigpu.py:323] [p0576]  step: count(603), step_time 1463.81, mean_step_time 1454.97, it/s 0.69
  3145. [0130 17:44:18 @multigpu.py:323] [p0115]  step: count(581), step_time 1418.59, mean_step_time 1432.75, it/s 0.7
  3146. [0130 17:44:18 @multigpu.py:323] [p0574]  step: count(597), step_time 1469.03, mean_step_time 1435.67, it/s 0.7
  3147. [0130 17:44:18 @multigpu.py:323] [p0576]  step: count(604), step_time 1420.29, mean_step_time 1457.42, it/s 0.69
  3148. [0130 17:44:20 @multigpu.py:323] [p0115]  step: count(582), step_time 1458.28, mean_step_time 1436.31, it/s 0.7
  3149. [0130 17:44:20 @multigpu.py:323] [p0574]  step: count(598), step_time 1442.36, mean_step_time 1434.58, it/s 0.7
  3150. [0130 17:44:20 @multigpu.py:323] [p0576]  step: count(605), step_time 1400.04, mean_step_time 1455.02, it/s 0.69
  3151. sending to address tcp://p0112:61216
  3152. ##### Sending to neptune:  online_score :  0.238823784126 , 1.4 #####
  3153. [u'online', 1.4]
  3154. receiving
  3155. [0130 17:44:21 @multigpu.py:323] [p0574]  step: count(599), step_time 1400.31, mean_step_time 1433.08, it/s 0.7
  3156. [0130 17:44:21 @multigpu.py:323] [p0115]  step: count(583), step_time 1498.62, mean_step_time 1438.1, it/s 0.7
  3157. [0130 17:44:21 @multigpu.py:323] [p0576]  step: count(606), step_time 1421.09, mean_step_time 1453.86, it/s 0.69
  3158. [0130 17:44:23 @multigpu.py:323] [p0115]  step: count(584), step_time 1412.91, mean_step_time 1438.97, it/s 0.69
  3159. [0130 17:44:23 @multigpu.py:323] [p0576]  step: count(607), step_time 1404.31, mean_step_time 1452.04, it/s 0.69
  3160. [0130 17:44:23 @multigpu.py:323] [p0574]  step: count(600), step_time 1464.53, mean_step_time 1435.97, it/s 0.7
  3161. sending debugging info...
  3162. sending to address tcp://p0112:61216
  3163. ##### Sending to neptune:  mean_delay :  0.239441586071 , 0.0 #####
  3164. ##### Sending to neptune:  max_delay :  0.239441586071 , -0.0 #####
  3165. sending to address tcp://p0112:61216
  3166. ##### Sending to neptune:  min_delay :  0.239441586071 , -0.0 #####
  3167. [u'delays', [0.0, -0.0, -0.0]]
  3168. receiving
  3169. ##### Sending to neptune:  cost :  0.239442152182 , -0.0125148463994 #####
  3170. ##### Sending to neptune:  policy_loss :  0.239442152182 , -0.49882593751 #####
  3171. sending to address tcp://p0112:61216
  3172. ##### Sending to neptune:  xentropy_loss :  0.239442152182 , -2.29190254211 #####
  3173. ##### Sending to neptune:  value_loss :  0.239442152182 , 1.18882787228 #####
  3174. ##### Sending to neptune:  advantage :  0.239442152182 , 0.00211322447285 #####
  3175. ##### Sending to neptune:  pred_reward :  0.239442152182 , 0.398575603962 #####
  3176. ##### Sending to neptune:  max_logit :  0.239442152182 , 0.181839182973 #####
  3177. [u'loss', -0.01251484639942646, -0.49882593750953674, -2.291902542114258, 1.1888278722763062, 0.002113224472850561, 0.3985756039619446, 0.18183918297290802]
  3178. receiving
  3179. ##### Sending to neptune:  active_relus :  0.239442588554 , 8800469.51 #####
  3180. ##### Sending to neptune:  dp_per_s :  0.239442588554 , 89.0259561405 #####
  3181. [u'other', 8800469.51, 89.02595614054479]
  3182. receiving
  3183. [0130 17:44:24 @multigpu.py:323] [p0115]  step: count(585), step_time 1430.45, mean_step_time 1440.72, it/s 0.69
  3184. [0130 17:44:24 @multigpu.py:323] [p0576]  step: count(608), step_time 1455.01, mean_step_time 1456.26, it/s 0.69
  3185. [0130 17:44:24 @multigpu.py:323] [p0574]  step: count(601), step_time 1447.28, mean_step_time 1436.48, it/s 0.7
  3186. [0130 17:44:25 @multigpu.py:323] [p0115]  step: count(586), step_time 1403.09, mean_step_time 1441.43, it/s 0.69
  3187. [0130 17:44:25 @multigpu.py:323] [p0574]  step: count(602), step_time 1389.81, mean_step_time 1434.76, it/s 0.7
  3188. [0130 17:44:25 @multigpu.py:323] [p0576]  step: count(609), step_time 1459.53, mean_step_time 1458.13, it/s 0.69
  3189. sending to address tcp://p0112:61216
  3190. ##### Sending to neptune:  online_score :  0.240421581931 , 1.6 #####
  3191. [u'online', 1.6]
  3192. receiving
  3193. [0130 17:44:27 @multigpu.py:323] [p0574]  step: count(603), step_time 1412.17, mean_step_time 1433.52, it/s 0.7
  3194. [0130 17:44:27 @multigpu.py:323] [p0115]  step: count(587), step_time 1454.29, mean_step_time 1439.97, it/s 0.69
  3195. [0130 17:44:27 @multigpu.py:323] [p0576]  step: count(610), step_time 1377.06, mean_step_time 1456.67, it/s 0.69
  3196. [0130 17:44:28 @multigpu.py:323] [p0115]  step: count(588), step_time 1469.67, mean_step_time 1436.24, it/s 0.7
  3197. [0130 17:44:28 @multigpu.py:323] [p0574]  step: count(604), step_time 1487.58, mean_step_time 1436.24, it/s 0.7
  3198. [0130 17:44:28 @multigpu.py:323] [p0576]  step: count(611), step_time 1462.63, mean_step_time 1436.31, it/s 0.7
  3199. [0130 17:44:30 @multigpu.py:323] [p0574]  step: count(605), step_time 1389.52, mean_step_time 1433.37, it/s 0.7
  3200. [0130 17:44:30 @multigpu.py:323] [p0115]  step: count(589), step_time 1390.94, mean_step_time 1435.99, it/s 0.7
  3201. [0130 17:44:30 @multigpu.py:323] [p0576]  step: count(612), step_time 1450.4, mean_step_time 1438.31, it/s 0.7
  3202. [0130 17:44:31 @multigpu.py:323] [p0576]  step: count(613), step_time 1381.31, mean_step_time 1436.49, it/s 0.7
  3203. [0130 17:44:31 @multigpu.py:323] [p0574]  step: count(606), step_time 1463.87, mean_step_time 1437.08, it/s 0.7
  3204. [0130 17:44:31 @multigpu.py:323] [p0115]  step: count(590), step_time 1475.51, mean_step_time 1438.7, it/s 0.7
  3205. [0130 17:44:33 @multigpu.py:323] [p0576]  step: count(614), step_time 1400.85, mean_step_time 1433.04, it/s 0.7
  3206. [0130 17:44:33 @multigpu.py:323] [p0115]  step: count(591), step_time 1427.56, mean_step_time 1438.81, it/s 0.7
  3207. [0130 17:44:33 @multigpu.py:323] [p0574]  step: count(607), step_time 1496.17, mean_step_time 1439.94, it/s 0.69
  3208. [0130 17:44:34 @multigpu.py:323] [p0576]  step: count(615), step_time 1415.65, mean_step_time 1430.97, it/s 0.7
  3209. [0130 17:44:34 @multigpu.py:323] [p0115]  step: count(592), step_time 1362.75, mean_step_time 1435.29, it/s 0.7
  3210. [0130 17:44:34 @multigpu.py:323] [p0574]  step: count(608), step_time 1415.51, mean_step_time 1436.29, it/s 0.7
  3211. [0130 17:44:35 @multigpu.py:323] [p0576]  step: count(616), step_time 1422.25, mean_step_time 1430.35, it/s 0.7
  3212. [0130 17:44:35 @multigpu.py:323] [p0115]  step: count(593), step_time 1428.81, mean_step_time 1436.52, it/s 0.7
  3213. [0130 17:44:36 @multigpu.py:323] [p0574]  step: count(609), step_time 1479.88, mean_step_time 1438.61, it/s 0.7
  3214. [0130 17:44:37 @multigpu.py:323] [p0576]  step: count(617), step_time 1405.63, mean_step_time 1428.7, it/s 0.7
  3215. [0130 17:44:37 @multigpu.py:323] [p0115]  step: count(594), step_time 1480.72, mean_step_time 1437.42, it/s 0.7
  3216. [0130 17:44:37 @multigpu.py:323] [p0574]  step: count(610), step_time 1439.04, mean_step_time 1437.52, it/s 0.7
  3217. [0130 17:44:38 @multigpu.py:323] [p0576]  step: count(618), step_time 1453.8, mean_step_time 1430.83, it/s 0.7
  3218. [0130 17:44:38 @multigpu.py:323] [p0115]  step: count(595), step_time 1421.83, mean_step_time 1437.11, it/s 0.7
  3219. [0130 17:44:38 @multigpu.py:323] [p0574]  step: count(611), step_time 1399.35, mean_step_time 1435.38, it/s 0.7
  3220. [0130 17:44:40 @multigpu.py:323] [p0576]  step: count(619), step_time 1407.74, mean_step_time 1429.63, it/s 0.7
  3221. [0130 17:44:40 @multigpu.py:323] [p0115]  step: count(596), step_time 1425.51, mean_step_time 1436.89, it/s 0.7
  3222. [0130 17:44:40 @multigpu.py:323] [p0574]  step: count(612), step_time 1478.83, mean_step_time 1438.14, it/s 0.7
  3223. [0130 17:44:41 @multigpu.py:323] [p0576]  step: count(620), step_time 1445.59, mean_step_time 1428.51, it/s 0.7
  3224. [0130 17:44:41 @multigpu.py:323] [p0115]  step: count(597), step_time 1423.78, mean_step_time 1437.72, it/s 0.7
  3225. [0130 17:44:41 @multigpu.py:323] [p0574]  step: count(613), step_time 1408.85, mean_step_time 1438.91, it/s 0.69
  3226. [0130 17:44:43 @multigpu.py:323] [p0576]  step: count(621), step_time 1422.57, mean_step_time 1427.57, it/s 0.7
  3227. [0130 17:44:43 @multigpu.py:323] [p0115]  step: count(598), step_time 1433.0, mean_step_time 1435.02, it/s 0.7
  3228. [0130 17:44:43 @multigpu.py:323] [p0574]  step: count(614), step_time 1403.31, mean_step_time 1437.91, it/s 0.7
  3229. [0130 17:44:44 @multigpu.py:323] [p0576]  step: count(622), step_time 1411.85, mean_step_time 1424.07, it/s 0.7
  3230. [0130 17:44:44 @multigpu.py:323] [p0115]  step: count(599), step_time 1452.46, mean_step_time 1437.85, it/s 0.7
  3231. [0130 17:44:44 @multigpu.py:323] [p0574]  step: count(615), step_time 1476.66, mean_step_time 1439.95, it/s 0.69
  3232. [0130 17:44:45 @multigpu.py:323] [p0576]  step: count(623), step_time 1407.37, mean_step_time 1421.25, it/s 0.7
  3233. [0130 17:44:45 @multigpu.py:323] [p0115]  step: count(600), step_time 1399.95, mean_step_time 1433.44, it/s 0.7
  3234. sending debugging info...
  3235. sending to address tcp://p0112:61216
  3236. ##### Sending to neptune:  mean_delay :  0.24579519444 , 0.0 #####
  3237. sending to address tcp://p0112:61216
  3238. ##### Sending to neptune:  max_delay :  0.24579519444 , -0.0 #####
  3239. ##### Sending to neptune:  min_delay :  0.24579519444 , -0.0 #####
  3240. [u'delays', [0.0, -0.0, -0.0]]
  3241. receiving
  3242. ##### Sending to neptune:  cost :  0.245795726644 , -0.0124126961455 #####
  3243. ##### Sending to neptune:  policy_loss :  0.245795726644 , -0.532102823257 #####
  3244. sending to address tcp://p0112:61216
  3245. ##### Sending to neptune:  xentropy_loss :  0.245795726644 , -2.29180145264 #####
  3246. ##### Sending to neptune:  value_loss :  0.245795726644 , 1.23507893085 #####
  3247. ##### Sending to neptune:  advantage :  0.245795726644 , 0.00228149211034 #####
  3248. ##### Sending to neptune:  pred_reward :  0.245795726644 , 0.404466867447 #####
  3249. ##### Sending to neptune:  max_logit :  0.245795726644 , 0.183123975992 #####
  3250. [u'loss', -0.01241269614547491, -0.5321028232574463, -2.2918014526367188, 1.2350789308547974, 0.0022814921103417873, 0.4044668674468994, 0.18312397599220276]
  3251. receiving
  3252. ##### Sending to neptune:  active_relus :  0.245796224409 , 8822338.12 #####
  3253. ##### Sending to neptune:  dp_per_s :  0.245796224409 , 89.0029782201 #####
  3254. [u'other', 8822338.12, 89.00297822013982]
  3255. receiving
  3256. [0130 17:44:46 @multigpu.py:323] [p0574]  step: count(616), step_time 1425.14, mean_step_time 1439.46, it/s 0.69
  3257. [0130 17:44:47 @multigpu.py:323] [p0576]  step: count(624), step_time 1462.41, mean_step_time 1423.35, it/s 0.7
  3258. [0130 17:44:47 @multigpu.py:323] [p0115]  step: count(601), step_time 1463.34, mean_step_time 1435.68, it/s 0.7
  3259. [0130 17:44:47 @multigpu.py:323] [p0574]  step: count(617), step_time 1418.75, mean_step_time 1436.95, it/s 0.7
  3260. [0130 17:44:48 @multigpu.py:323] [p0576]  step: count(625), step_time 1426.03, mean_step_time 1424.65, it/s 0.7
  3261. [0130 17:44:48 @multigpu.py:323] [p0115]  step: count(602), step_time 1403.93, mean_step_time 1432.96, it/s 0.7
  3262. [0130 17:44:48 @multigpu.py:323] [p0574]  step: count(618), step_time 1393.44, mean_step_time 1434.5, it/s 0.7
  3263. sending to address tcp://p0112:61216
  3264. ##### Sending to neptune:  online_score :  0.246657715506 , 1.4 #####
  3265. [u'online', 1.4]
  3266. receiving
  3267. sending to address tcp://p0112:61216
  3268. ##### Sending to neptune:  online_score :  0.246721496052 , 1.6 #####
  3269. [u'online', 1.6]
  3270. receiving
  3271. [0130 17:44:50 @multigpu.py:323] [p0576]  step: count(626), step_time 1402.71, mean_step_time 1423.73, it/s 0.7
  3272. [0130 17:44:50 @multigpu.py:323] [p0115]  step: count(603), step_time 1429.02, mean_step_time 1429.48, it/s 0.7
  3273. [0130 17:44:50 @multigpu.py:323] [p0574]  step: count(619), step_time 1492.01, mean_step_time 1439.09, it/s 0.69
  3274. sending to address tcp://p0112:61216
  3275. ##### Sending to neptune:  online_score :  0.247186529703 , 0.6 #####
  3276. [u'online', 0.6]
  3277. receiving
  3278. [0130 17:44:51 @multigpu.py:323] [p0576]  step: count(627), step_time 1407.93, mean_step_time 1423.92, it/s 0.7
  3279. [0130 17:44:51 @multigpu.py:323] [p0115]  step: count(604), step_time 1412.04, mean_step_time 1429.43, it/s 0.7
  3280. [0130 17:44:51 @multigpu.py:323] [p0574]  step: count(620), step_time 1405.33, mean_step_time 1436.13, it/s 0.7
  3281. [0130 17:44:52 @multigpu.py:323] [p0576]  step: count(628), step_time 1390.03, mean_step_time 1420.67, it/s 0.7
  3282. [0130 17:44:53 @multigpu.py:323] [p0115]  step: count(605), step_time 1426.99, mean_step_time 1429.26, it/s 0.7
  3283. [0130 17:44:53 @multigpu.py:323] [p0574]  step: count(621), step_time 1503.47, mean_step_time 1438.93, it/s 0.69
  3284. [0130 17:44:54 @multigpu.py:323] [p0576]  step: count(629), step_time 1378.84, mean_step_time 1416.63, it/s 0.71
  3285. [0130 17:44:54 @multigpu.py:323] [p0115]  step: count(606), step_time 1481.85, mean_step_time 1433.2, it/s 0.7
  3286. [0130 17:44:54 @multigpu.py:323] [p0574]  step: count(622), step_time 1415.95, mean_step_time 1440.24, it/s 0.69
  3287. [0130 17:44:55 @multigpu.py:323] [p0576]  step: count(630), step_time 1494.85, mean_step_time 1422.52, it/s 0.7
  3288. [0130 17:44:55 @multigpu.py:323] [p0115]  step: count(607), step_time 1399.97, mean_step_time 1430.48, it/s 0.7
  3289. [0130 17:44:56 @multigpu.py:323] [p0574]  step: count(623), step_time 1419.95, mean_step_time 1440.63, it/s 0.69
  3290. [0130 17:44:57 @multigpu.py:323] [p0115]  step: count(608), step_time 1606.73, mean_step_time 1437.34, it/s 0.7
  3291. [0130 17:44:57 @multigpu.py:323] [p0574]  step: count(624), step_time 1424.07, mean_step_time 1437.46, it/s 0.7
  3292. [0130 17:44:57 @multigpu.py:323] [p0576]  step: count(631), step_time 1761.94, mean_step_time 1437.49, it/s 0.7
  3293. [0130 17:44:59 @multigpu.py:323] [p0576]  step: count(632), step_time 1410.12, mean_step_time 1435.47, it/s 0.7
  3294. [0130 17:44:59 @multigpu.py:323] [p0115]  step: count(609), step_time 1436.63, mean_step_time 1439.62, it/s 0.69
  3295. [0130 17:44:59 @multigpu.py:323] [p0574]  step: count(625), step_time 1486.65, mean_step_time 1442.31, it/s 0.69
  3296. [0130 17:45:00 @multigpu.py:323] [p0576]  step: count(633), step_time 1481.32, mean_step_time 1440.47, it/s 0.69
  3297. [0130 17:45:00 @multigpu.py:323] [p0115]  step: count(610), step_time 1477.83, mean_step_time 1439.74, it/s 0.69
  3298. [0130 17:45:00 @multigpu.py:323] [p0574]  step: count(626), step_time 1486.27, mean_step_time 1443.43, it/s 0.69
  3299. [0130 17:45:01 @multigpu.py:323] [p0115]  step: count(611), step_time 1405.17, mean_step_time 1438.62, it/s 0.7
  3300. [0130 17:45:01 @multigpu.py:323] [p0576]  step: count(634), step_time 1428.84, mean_step_time 1441.87, it/s 0.69
  3301. [0130 17:45:01 @multigpu.py:323] [p0574]  step: count(627), step_time 1387.93, mean_step_time 1438.02, it/s 0.7
  3302. [0130 17:45:03 @multigpu.py:323] [p0115]  step: count(612), step_time 1414.92, mean_step_time 1441.22, it/s 0.69
  3303. [0130 17:45:03 @multigpu.py:323] [p0574]  step: count(628), step_time 1438.6, mean_step_time 1439.17, it/s 0.69
  3304. [0130 17:45:03 @multigpu.py:323] [p0576]  step: count(635), step_time 1515.42, mean_step_time 1446.86, it/s 0.69
  3305. [0130 17:45:04 @multigpu.py:323] [p0115]  step: count(613), step_time 1465.02, mean_step_time 1443.03, it/s 0.69
  3306. [0130 17:45:04 @multigpu.py:323] [p0576]  step: count(636), step_time 1393.03, mean_step_time 1445.4, it/s 0.69
  3307. [0130 17:45:04 @multigpu.py:323] [p0574]  step: count(629), step_time 1468.57, mean_step_time 1438.61, it/s 0.7
  3308. [0130 17:45:06 @multigpu.py:323] [p0115]  step: count(614), step_time 1453.47, mean_step_time 1441.67, it/s 0.69
  3309. [0130 17:45:06 @multigpu.py:323] [p0576]  step: count(637), step_time 1430.3, mean_step_time 1446.63, it/s 0.69
  3310. [0130 17:45:06 @multigpu.py:323] [p0574]  step: count(630), step_time 1414.37, mean_step_time 1437.38, it/s 0.7
  3311. [0130 17:45:07 @multigpu.py:323] [p0574]  step: count(631), step_time 1374.32, mean_step_time 1436.12, it/s 0.7
  3312. [0130 17:45:07 @multigpu.py:323] [p0576]  step: count(638), step_time 1400.2, mean_step_time 1443.95, it/s 0.69
  3313. [0130 17:45:07 @multigpu.py:323] [p0115]  step: count(615), step_time 1432.96, mean_step_time 1442.23, it/s 0.69
  3314. [0130 17:45:09 @multigpu.py:323] [p0576]  step: count(639), step_time 1464.58, mean_step_time 1446.8, it/s 0.69
  3315. [0130 17:45:09 @multigpu.py:323] [p0574]  step: count(632), step_time 1482.66, mean_step_time 1436.32, it/s 0.7
  3316. [0130 17:45:09 @multigpu.py:323] [p0115]  step: count(616), step_time 1473.6, mean_step_time 1444.63, it/s 0.69
  3317. [0130 17:45:10 @multigpu.py:323] [p0574]  step: count(633), step_time 1401.24, mean_step_time 1435.94, it/s 0.7
  3318. [0130 17:45:10 @multigpu.py:323] [p0115]  step: count(617), step_time 1406.4, mean_step_time 1443.76, it/s 0.69
  3319. [0130 17:45:10 @multigpu.py:323] [p0576]  step: count(640), step_time 1458.48, mean_step_time 1447.44, it/s 0.69
  3320. [0130 17:45:11 @multigpu.py:323] [p0115]  step: count(618), step_time 1410.03, mean_step_time 1442.61, it/s 0.69
  3321. [0130 17:45:11 @multigpu.py:323] [p0574]  step: count(634), step_time 1454.65, mean_step_time 1438.5, it/s 0.7
  3322. [0130 17:45:12 @multigpu.py:323] [p0576]  step: count(641), step_time 1431.78, mean_step_time 1447.9, it/s 0.69
  3323. [0130 17:45:13 @multigpu.py:323] [p0115]  step: count(619), step_time 1410.05, mean_step_time 1440.49, it/s 0.69
  3324. [0130 17:45:13 @multigpu.py:323] [p0576]  step: count(642), step_time 1456.59, mean_step_time 1450.14, it/s 0.69
  3325. [0130 17:45:13 @multigpu.py:323] [p0574]  step: count(635), step_time 1477.97, mean_step_time 1438.57, it/s 0.7
  3326. [0130 17:45:14 @multigpu.py:323] [p0115]  step: count(620), step_time 1455.96, mean_step_time 1443.29, it/s 0.69
  3327. [0130 17:45:14 @multigpu.py:323] [p0576]  step: count(643), step_time 1389.1, mean_step_time 1449.22, it/s 0.69
  3328. [0130 17:45:14 @multigpu.py:323] [p0574]  step: count(636), step_time 1431.19, mean_step_time 1438.87, it/s 0.69
  3329. sending to address tcp://p0112:61216
  3330. ##### Sending to neptune:  online_score :  0.25393662552 , 1.5 #####
  3331. [u'online', 1.5]
  3332. ##### Sending to neptune:  active_workers :  0.253936661614 , 3 #####
  3333. receiving
  3334. sending to address tcp://p0112:61216
  3335. ##### Sending to neptune:  online_score :  0.253976651099 , 1.5 #####
  3336. [u'online', 1.5]
  3337. receiving
  3338. [0130 17:45:16 @multigpu.py:323] [p0115]  step: count(621), step_time 1374.5, mean_step_time 1438.85, it/s 0.69
  3339. [0130 17:45:16 @multigpu.py:323] [p0576]  step: count(644), step_time 1412.07, mean_step_time 1446.71, it/s 0.69
  3340. [0130 17:45:16 @multigpu.py:323] [p0574]  step: count(637), step_time 1401.14, mean_step_time 1437.99, it/s 0.7
  3341. [0130 17:45:17 @multigpu.py:323] [p0115]  step: count(622), step_time 1455.7, mean_step_time 1441.44, it/s 0.69
  3342. [0130 17:45:17 @multigpu.py:323] [p0574]  step: count(638), step_time 1388.99, mean_step_time 1437.77, it/s 0.7
  3343. [0130 17:45:17 @multigpu.py:323] [p0576]  step: count(645), step_time 1450.71, mean_step_time 1447.94, it/s 0.69
  3344. [0130 17:45:19 @multigpu.py:323] [p0576]  step: count(646), step_time 1429.79, mean_step_time 1449.3, it/s 0.69
  3345. [0130 17:45:19 @multigpu.py:323] [p0115]  step: count(623), step_time 1483.14, mean_step_time 1444.15, it/s 0.69
  3346. [0130 17:45:19 @multigpu.py:323] [p0574]  step: count(639), step_time 1500.31, mean_step_time 1438.18, it/s 0.7
  3347. sending to address tcp://p0112:61216
  3348. ##### Sending to neptune:  online_score :  0.255228105783 , 1.5 #####
  3349. [u'online', 1.5]
  3350. receiving
  3351. [0130 17:45:20 @multigpu.py:323] [p0115]  step: count(624), step_time 1400.48, mean_step_time 1443.57, it/s 0.69
  3352. [0130 17:45:20 @multigpu.py:323] [p0576]  step: count(647), step_time 1429.17, mean_step_time 1450.36, it/s 0.69
  3353. [0130 17:45:20 @multigpu.py:323] [p0574]  step: count(640), step_time 1431.48, mean_step_time 1439.49, it/s 0.69
  3354. [0130 17:45:21 @multigpu.py:323] [p0115]  step: count(625), step_time 1438.56, mean_step_time 1444.15, it/s 0.69
  3355. [0130 17:45:22 @multigpu.py:323] [p0574]  step: count(641), step_time 1377.16, mean_step_time 1433.17, it/s 0.7
  3356. [0130 17:45:22 @multigpu.py:323] [p0576]  step: count(648), step_time 1458.03, mean_step_time 1453.76, it/s 0.69
  3357. [0130 17:45:23 @multigpu.py:323] [p0574]  step: count(642), step_time 1409.73, mean_step_time 1432.86, it/s 0.7
  3358. [0130 17:45:23 @multigpu.py:323] [p0115]  step: count(626), step_time 1456.32, mean_step_time 1442.87, it/s 0.69
  3359. [0130 17:45:23 @multigpu.py:323] [p0576]  step: count(649), step_time 1492.76, mean_step_time 1459.45, it/s 0.69
  3360. [0130 17:45:24 @multigpu.py:323] [p0574]  step: count(643), step_time 1467.67, mean_step_time 1435.25, it/s 0.7
  3361. [0130 17:45:24 @multigpu.py:323] [p0115]  step: count(627), step_time 1441.65, mean_step_time 1444.96, it/s 0.69
  3362. [0130 17:45:24 @multigpu.py:323] [p0576]  step: count(650), step_time 1411.33, mean_step_time 1455.28, it/s 0.69
  3363. [0130 17:45:26 @multigpu.py:323] [p0115]  step: count(628), step_time 1539.64, mean_step_time 1441.6, it/s 0.69
  3364. [0130 17:45:26 @multigpu.py:323] [p0576]  step: count(651), step_time 1488.15, mean_step_time 1441.59, it/s 0.69
  3365. [0130 17:45:26 @multigpu.py:323] [p0574]  step: count(644), step_time 1551.55, mean_step_time 1441.62, it/s 0.69
  3366. [0130 17:45:27 @multigpu.py:323] [p0576]  step: count(652), step_time 1392.0, mean_step_time 1440.68, it/s 0.69
  3367. [0130 17:45:27 @multigpu.py:323] [p0115]  step: count(629), step_time 1405.88, mean_step_time 1440.06, it/s 0.69
  3368. [0130 17:45:27 @multigpu.py:323] [p0574]  step: count(645), step_time 1471.01, mean_step_time 1440.84, it/s 0.69
  3369. [0130 17:45:29 @multigpu.py:323] [p0576]  step: count(653), step_time 1416.31, mean_step_time 1437.43, it/s 0.7
  3370. [0130 17:45:29 @multigpu.py:323] [p0115]  step: count(630), step_time 1429.28, mean_step_time 1437.64, it/s 0.7
  3371. [0130 17:45:29 @multigpu.py:323] [p0574]  step: count(646), step_time 1418.72, mean_step_time 1437.46, it/s 0.7
  3372. [0130 17:45:30 @multigpu.py:323] [p0576]  step: count(654), step_time 1443.74, mean_step_time 1438.18, it/s 0.7
  3373. [0130 17:45:30 @multigpu.py:323] [p0115]  step: count(631), step_time 1452.88, mean_step_time 1440.02, it/s 0.69
  3374. [0130 17:45:30 @multigpu.py:323] [p0574]  step: count(647), step_time 1431.32, mean_step_time 1439.63, it/s 0.69
  3375. [0130 17:45:32 @multigpu.py:323] [p0576]  step: count(655), step_time 1400.57, mean_step_time 1432.44, it/s 0.7
  3376. [0130 17:45:32 @multigpu.py:323] [p0115]  step: count(632), step_time 1430.96, mean_step_time 1440.82, it/s 0.69
  3377. [0130 17:45:32 @multigpu.py:323] [p0574]  step: count(648), step_time 1474.59, mean_step_time 1441.43, it/s 0.69
  3378. [0130 17:45:33 @multigpu.py:323] [p0576]  step: count(656), step_time 1432.96, mean_step_time 1434.43, it/s 0.7
  3379. [0130 17:45:33 @multigpu.py:323] [p0115]  step: count(633), step_time 1431.99, mean_step_time 1439.17, it/s 0.69
  3380. [0130 17:45:33 @multigpu.py:323] [p0574]  step: count(649), step_time 1383.33, mean_step_time 1437.17, it/s 0.7
  3381. [0130 17:45:34 @multigpu.py:323] [p0576]  step: count(657), step_time 1400.38, mean_step_time 1432.94, it/s 0.7
  3382. [0130 17:45:35 @multigpu.py:323] [p0574]  step: count(650), step_time 1413.86, mean_step_time 1437.14, it/s 0.7
  3383. [0130 17:45:35 @multigpu.py:323] [p0115]  step: count(634), step_time 1454.74, mean_step_time 1439.24, it/s 0.69
  3384. [0130 17:45:36 @multigpu.py:323] [p0576]  step: count(658), step_time 1378.2, mean_step_time 1431.84, it/s 0.7
  3385. [0130 17:45:36 @multigpu.py:323] [p0574]  step: count(651), step_time 1426.33, mean_step_time 1439.74, it/s 0.69
  3386. [0130 17:45:36 @multigpu.py:323] [p0115]  step: count(635), step_time 1506.58, mean_step_time 1442.92, it/s 0.69
  3387. [0130 17:45:37 @multigpu.py:323] [p0576]  step: count(659), step_time 1411.7, mean_step_time 1429.19, it/s 0.7
  3388. [0130 17:45:37 @multigpu.py:323] [p0574]  step: count(652), step_time 1434.51, mean_step_time 1437.34, it/s 0.7
  3389. [0130 17:45:38 @multigpu.py:323] [p0115]  step: count(636), step_time 1487.17, mean_step_time 1443.59, it/s 0.69
  3390. [0130 17:45:39 @multigpu.py:323] [p0576]  step: count(660), step_time 1500.11, mean_step_time 1431.27, it/s 0.7
  3391. [0130 17:45:39 @multigpu.py:323] [p0574]  step: count(653), step_time 1401.87, mean_step_time 1437.37, it/s 0.7
  3392. [0130 17:45:39 @multigpu.py:323] [p0115]  step: count(637), step_time 1403.42, mean_step_time 1443.45, it/s 0.69
  3393. sending to address tcp://p0112:61216
  3394. ##### Sending to neptune:  online_score :  0.260792402493 , 1.3 #####
  3395. [u'online', 1.3]
  3396. receiving
  3397. [0130 17:45:40 @multigpu.py:323] [p0576]  step: count(661), step_time 1437.46, mean_step_time 1431.56, it/s 0.7
  3398. [0130 17:45:40 @multigpu.py:323] [p0574]  step: count(654), step_time 1417.83, mean_step_time 1435.53, it/s 0.7
  3399. [0130 17:45:40 @multigpu.py:323] [p0115]  step: count(638), step_time 1405.06, mean_step_time 1443.2, it/s 0.69
  3400. [0130 17:45:42 @multigpu.py:323] [p0576]  step: count(662), step_time 1407.57, mean_step_time 1429.11, it/s 0.7
  3401. [0130 17:45:42 @multigpu.py:323] [p0574]  step: count(655), step_time 1442.15, mean_step_time 1433.74, it/s 0.7
  3402. [0130 17:45:42 @multigpu.py:323] [p0115]  step: count(639), step_time 1439.66, mean_step_time 1444.68, it/s 0.69
  3403. sending to address tcp://p0112:61216
  3404. ##### Sending to neptune:  online_score :  0.261552384694 , 1.6 #####
  3405. [u'online', 1.6]
  3406. receiving
  3407. [0130 17:45:43 @multigpu.py:323] [p0576]  step: count(663), step_time 1436.11, mean_step_time 1431.46, it/s 0.7
  3408. [0130 17:45:43 @multigpu.py:323] [p0574]  step: count(656), step_time 1442.82, mean_step_time 1434.32, it/s 0.7
  3409. [0130 17:45:43 @multigpu.py:323] [p0115]  step: count(640), step_time 1440.04, mean_step_time 1443.88, it/s 0.69
  3410. [0130 17:45:44 @multigpu.py:323] [p0576]  step: count(664), step_time 1388.96, mean_step_time 1430.3, it/s 0.7
  3411. [0130 17:45:45 @multigpu.py:323] [p0574]  step: count(657), step_time 1426.29, mean_step_time 1435.58, it/s 0.7
  3412. [0130 17:45:45 @multigpu.py:323] [p0115]  step: count(641), step_time 1479.51, mean_step_time 1449.13, it/s 0.69
  3413. [0130 17:45:46 @multigpu.py:323] [p0576]  step: count(665), step_time 1508.77, mean_step_time 1433.2, it/s 0.7
  3414. [0130 17:45:46 @multigpu.py:323] [p0574]  step: count(658), step_time 1395.82, mean_step_time 1435.92, it/s 0.7
  3415. [0130 17:45:46 @multigpu.py:323] [p0115]  step: count(642), step_time 1406.01, mean_step_time 1446.65, it/s 0.69
  3416. [0130 17:45:47 @multigpu.py:323] [p0576]  step: count(666), step_time 1405.08, mean_step_time 1431.97, it/s 0.7
  3417. [0130 17:45:47 @multigpu.py:323] [p0574]  step: count(659), step_time 1418.44, mean_step_time 1431.82, it/s 0.7
  3418. [0130 17:45:48 @multigpu.py:323] [p0115]  step: count(643), step_time 1422.24, mean_step_time 1443.6, it/s 0.69
  3419. sending to address tcp://p0112:61216
  3420. ##### Sending to neptune:  online_score :  0.263147509164 , 1.8 #####
  3421. [u'online', 1.8]
  3422. receiving
  3423. [0130 17:45:49 @multigpu.py:323] [p0576]  step: count(667), step_time 1399.86, mean_step_time 1430.5, it/s 0.7
  3424. [0130 17:45:49 @multigpu.py:323] [p0574]  step: count(660), step_time 1423.0, mean_step_time 1431.4, it/s 0.7
  3425. [0130 17:45:49 @multigpu.py:323] [p0115]  step: count(644), step_time 1418.9, mean_step_time 1444.52, it/s 0.69
  3426. [0130 17:45:50 @multigpu.py:323] [p0576]  step: count(668), step_time 1365.71, mean_step_time 1425.89, it/s 0.7
  3427. [0130 17:45:50 @multigpu.py:323] [p0574]  step: count(661), step_time 1455.74, mean_step_time 1435.33, it/s 0.7
  3428. [0130 17:45:50 @multigpu.py:323] [p0115]  step: count(645), step_time 1434.67, mean_step_time 1444.33, it/s 0.69
  3429. [0130 17:45:52 @multigpu.py:323] [p0576]  step: count(669), step_time 1467.99, mean_step_time 1424.65, it/s 0.7
  3430. [0130 17:45:52 @multigpu.py:323] [p0574]  step: count(662), step_time 1431.18, mean_step_time 1436.4, it/s 0.7
  3431. [0130 17:45:52 @multigpu.py:323] [p0115]  step: count(646), step_time 1457.55, mean_step_time 1444.39, it/s 0.69
  3432. [0130 17:45:53 @multigpu.py:323] [p0576]  step: count(670), step_time 1433.2, mean_step_time 1425.74, it/s 0.7
  3433. [0130 17:45:53 @multigpu.py:323] [p0574]  step: count(663), step_time 1384.25, mean_step_time 1432.23, it/s 0.7
  3434. [0130 17:45:53 @multigpu.py:323] [p0115]  step: count(647), step_time 1400.16, mean_step_time 1442.32, it/s 0.69
  3435. [0130 17:45:55 @multigpu.py:323] [p0115]  step: count(648), step_time 1497.96, mean_step_time 1440.23, it/s 0.69
  3436. [0130 17:45:55 @multigpu.py:323] [p0574]  step: count(664), step_time 1712.16, mean_step_time 1440.26, it/s 0.69
  3437. [0130 17:45:55 @multigpu.py:323] [p0576]  step: count(671), step_time 1778.23, mean_step_time 1440.25, it/s 0.69
  3438. [0130 17:45:56 @multigpu.py:323] [p0574]  step: count(665), step_time 1379.79, mean_step_time 1435.7, it/s 0.7
  3439. [0130 17:45:56 @multigpu.py:323] [p0576]  step: count(672), step_time 1433.5, mean_step_time 1442.32, it/s 0.69
  3440. [0130 17:45:56 @multigpu.py:323] [p0115]  step: count(649), step_time 1457.17, mean_step_time 1442.8, it/s 0.69
  3441. [0130 17:45:58 @multigpu.py:323] [p0576]  step: count(673), step_time 1417.96, mean_step_time 1442.4, it/s 0.69
  3442. [0130 17:45:58 @multigpu.py:323] [p0574]  step: count(666), step_time 1481.13, mean_step_time 1438.82, it/s 0.7
  3443. [0130 17:45:58 @multigpu.py:323] [p0115]  step: count(650), step_time 1437.6, mean_step_time 1443.21, it/s 0.69
  3444. [0130 17:45:59 @multigpu.py:323] [p0574]  step: count(667), step_time 1362.51, mean_step_time 1435.38, it/s 0.7
  3445. [0130 17:45:59 @multigpu.py:323] [p0576]  step: count(674), step_time 1432.19, mean_step_time 1441.83, it/s 0.69
  3446. [0130 17:45:59 @multigpu.py:323] [p0115]  step: count(651), step_time 1409.93, mean_step_time 1441.07, it/s 0.69
  3447. [0130 17:46:00 @multigpu.py:323] [p0574]  step: count(668), step_time 1451.17, mean_step_time 1434.21, it/s 0.7
  3448. [0130 17:46:00 @multigpu.py:323] [p0115]  step: count(652), step_time 1423.18, mean_step_time 1440.68, it/s 0.69
  3449. [0130 17:46:01 @multigpu.py:323] [p0576]  step: count(675), step_time 1480.5, mean_step_time 1445.82, it/s 0.69
  3450. [0130 17:46:02 @multigpu.py:323] [p0574]  step: count(669), step_time 1377.37, mean_step_time 1433.91, it/s 0.7
  3451. [0130 17:46:02 @multigpu.py:323] [p0576]  step: count(676), step_time 1375.98, mean_step_time 1442.97, it/s 0.69
  3452. [0130 17:46:02 @multigpu.py:323] [p0115]  step: count(653), step_time 1451.27, mean_step_time 1441.64, it/s 0.69
  3453. [0130 17:46:03 @multigpu.py:323] [p0574]  step: count(670), step_time 1407.84, mean_step_time 1433.61, it/s 0.7
  3454. [0130 17:46:03 @multigpu.py:323] [p0576]  step: count(677), step_time 1418.6, mean_step_time 1443.88, it/s 0.69
  3455. [0130 17:46:03 @multigpu.py:323] [p0115]  step: count(654), step_time 1465.6, mean_step_time 1442.18, it/s 0.69
  3456. [0130 17:46:05 @multigpu.py:323] [p0574]  step: count(671), step_time 1388.14, mean_step_time 1431.7, it/s 0.7
  3457. [0130 17:46:05 @multigpu.py:323] [p0576]  step: count(678), step_time 1384.74, mean_step_time 1444.21, it/s 0.69
  3458. [0130 17:46:05 @multigpu.py:323] [p0115]  step: count(655), step_time 1431.25, mean_step_time 1438.42, it/s 0.7
  3459. [0130 17:46:06 @multigpu.py:323] [p0574]  step: count(672), step_time 1407.39, mean_step_time 1430.34, it/s 0.7
  3460. [0130 17:46:06 @multigpu.py:323] [p0576]  step: count(679), step_time 1490.79, mean_step_time 1448.17, it/s 0.69
  3461. [0130 17:46:06 @multigpu.py:323] [p0115]  step: count(656), step_time 1417.72, mean_step_time 1434.94, it/s 0.7
  3462. [0130 17:46:07 @multigpu.py:323] [p0574]  step: count(673), step_time 1434.03, mean_step_time 1431.95, it/s 0.7
  3463. [0130 17:46:08 @multigpu.py:323] [p0576]  step: count(680), step_time 1383.0, mean_step_time 1442.31, it/s 0.69
  3464. [0130 17:46:08 @multigpu.py:323] [p0115]  step: count(657), step_time 1430.33, mean_step_time 1436.29, it/s 0.7
  3465. [0130 17:46:09 @multigpu.py:323] [p0574]  step: count(674), step_time 1434.18, mean_step_time 1432.77, it/s 0.7
  3466. [0130 17:46:09 @multigpu.py:323] [p0576]  step: count(681), step_time 1426.53, mean_step_time 1441.76, it/s 0.69
  3467. [0130 17:46:09 @multigpu.py:323] [p0115]  step: count(658), step_time 1511.28, mean_step_time 1441.6, it/s 0.69
  3468. [0130 17:46:10 @multigpu.py:323] [p0574]  step: count(675), step_time 1409.21, mean_step_time 1431.12, it/s 0.7
  3469. [0130 17:46:10 @multigpu.py:323] [p0576]  step: count(682), step_time 1378.39, mean_step_time 1440.31, it/s 0.69
  3470. [0130 17:46:11 @multigpu.py:323] [p0115]  step: count(659), step_time 1437.91, mean_step_time 1441.51, it/s 0.69
  3471. sending to address tcp://p0112:61216
  3472. ##### Sending to neptune:  online_score :  0.269616084165 , 1.4 #####
  3473. [u'online', 1.4]
  3474. receiving
  3475. [0130 17:46:12 @multigpu.py:323] [p0574]  step: count(676), step_time 1388.14, mean_step_time 1428.39, it/s 0.7
  3476. [0130 17:46:12 @multigpu.py:323] [p0576]  step: count(683), step_time 1416.55, mean_step_time 1439.33, it/s 0.69
  3477. sending to address tcp://p0112:61216
  3478. ##### Sending to neptune:  online_score :  0.269859721661 , 1.8 #####
  3479. [u'online', 1.8]
  3480. receiving
  3481. [0130 17:46:12 @multigpu.py:323] [p0115]  step: count(660), step_time 1474.67, mean_step_time 1443.25, it/s 0.69
  3482. [0130 17:46:13 @multigpu.py:323] [p0574]  step: count(677), step_time 1442.56, mean_step_time 1429.2, it/s 0.7
  3483. [0130 17:46:13 @multigpu.py:323] [p0576]  step: count(684), step_time 1408.24, mean_step_time 1440.29, it/s 0.69
  3484. [0130 17:46:14 @multigpu.py:323] [p0115]  step: count(661), step_time 1416.46, mean_step_time 1440.09, it/s 0.69
  3485. [0130 17:46:15 @multigpu.py:323] [p0574]  step: count(678), step_time 1410.63, mean_step_time 1429.94, it/s 0.7
  3486. [0130 17:46:15 @multigpu.py:323] [p0576]  step: count(685), step_time 1397.51, mean_step_time 1434.73, it/s 0.7
  3487. [0130 17:46:15 @multigpu.py:323] [p0115]  step: count(662), step_time 1455.37, mean_step_time 1442.56, it/s 0.69
  3488. sending to address tcp://p0112:61216
  3489. ##### Sending to neptune:  online_score :  0.270699454414 , 1.3 #####
  3490. [u'online', 1.3]
  3491. ##### Sending to neptune:  active_workers :  0.270699488322 , 3 #####
  3492. receiving
  3493. [0130 17:46:16 @multigpu.py:323] [p0574]  step: count(679), step_time 1416.51, mean_step_time 1429.85, it/s 0.7
  3494. [0130 17:46:16 @multigpu.py:323] [p0576]  step: count(686), step_time 1431.12, mean_step_time 1436.03, it/s 0.7
  3495. [0130 17:46:16 @multigpu.py:323] [p0115]  step: count(663), step_time 1399.97, mean_step_time 1441.45, it/s 0.69
  3496. [0130 17:46:17 @multigpu.py:323] [p0574]  step: count(680), step_time 1418.6, mean_step_time 1429.63, it/s 0.7
  3497. [0130 17:46:18 @multigpu.py:323] [p0576]  step: count(687), step_time 1467.33, mean_step_time 1439.4, it/s 0.69
  3498. [0130 17:46:18 @multigpu.py:323] [p0115]  step: count(664), step_time 1441.72, mean_step_time 1442.59, it/s 0.69
  3499. [0130 17:46:19 @multigpu.py:323] [p0574]  step: count(681), step_time 1434.2, mean_step_time 1428.55, it/s 0.7
  3500. [0130 17:46:19 @multigpu.py:323] [p0576]  step: count(688), step_time 1412.21, mean_step_time 1441.73, it/s 0.69
  3501. [0130 17:46:19 @multigpu.py:323] [p0115]  step: count(665), step_time 1450.61, mean_step_time 1443.39, it/s 0.69
  3502. [0130 17:46:20 @multigpu.py:323] [p0574]  step: count(682), step_time 1371.35, mean_step_time 1425.56, it/s 0.7
  3503. [0130 17:46:20 @multigpu.py:323] [p0576]  step: count(689), step_time 1424.95, mean_step_time 1439.58, it/s 0.69
  3504. [0130 17:46:21 @multigpu.py:323] [p0115]  step: count(666), step_time 1402.55, mean_step_time 1440.64, it/s 0.69
  3505. [0130 17:46:22 @multigpu.py:323] [p0574]  step: count(683), step_time 1467.57, mean_step_time 1429.72, it/s 0.7
  3506. [0130 17:46:22 @multigpu.py:323] [p0576]  step: count(690), step_time 1417.75, mean_step_time 1438.8, it/s 0.7
  3507. [0130 17:46:22 @multigpu.py:323] [p0115]  step: count(667), step_time 1430.56, mean_step_time 1442.16, it/s 0.69
  3508. [0130 17:46:24 @multigpu.py:323] [p0576]  step: count(691), step_time 1760.88, mean_step_time 1437.94, it/s 0.7
  3509. [0130 17:46:24 @multigpu.py:323] [p0574]  step: count(684), step_time 1876.98, mean_step_time 1437.97, it/s 0.7
  3510. [0130 17:46:24 @multigpu.py:323] [p0115]  step: count(668), step_time 1414.87, mean_step_time 1438.0, it/s 0.7
  3511. [0130 17:46:25 @multigpu.py:323] [p0576]  step: count(692), step_time 1412.65, mean_step_time 1436.89, it/s 0.7
  3512. [0130 17:46:25 @multigpu.py:323] [p0574]  step: count(685), step_time 1433.75, mean_step_time 1440.66, it/s 0.69
  3513. [0130 17:46:25 @multigpu.py:323] [p0115]  step: count(669), step_time 1442.29, mean_step_time 1437.26, it/s 0.7
  3514. [0130 17:46:26 @multigpu.py:323] [p0576]  step: count(693), step_time 1410.89, mean_step_time 1436.54, it/s 0.7
  3515. [0130 17:46:26 @multigpu.py:323] [p0115]  step: count(670), step_time 1392.77, mean_step_time 1435.02, it/s 0.7
  3516. [0130 17:46:26 @multigpu.py:323] [p0574]  step: count(686), step_time 1417.46, mean_step_time 1437.48, it/s 0.7
  3517. [0130 17:46:28 @multigpu.py:323] [p0115]  step: count(671), step_time 1422.11, mean_step_time 1435.62, it/s 0.7
  3518. [0130 17:46:28 @multigpu.py:323] [p0574]  step: count(687), step_time 1428.69, mean_step_time 1440.79, it/s 0.69
  3519. [0130 17:46:28 @multigpu.py:323] [p0576]  step: count(694), step_time 1488.4, mean_step_time 1439.35, it/s 0.69
  3520. [0130 17:46:29 @multigpu.py:323] [p0576]  step: count(695), step_time 1400.05, mean_step_time 1435.33, it/s 0.7
  3521. [0130 17:46:29 @multigpu.py:323] [p0115]  step: count(672), step_time 1488.86, mean_step_time 1438.91, it/s 0.69
  3522. [0130 17:46:29 @multigpu.py:323] [p0574]  step: count(688), step_time 1470.6, mean_step_time 1441.76, it/s 0.69
  3523. [0130 17:46:31 @multigpu.py:323] [p0574]  step: count(689), step_time 1413.16, mean_step_time 1443.55, it/s 0.69
  3524. [0130 17:46:31 @multigpu.py:323] [p0115]  step: count(673), step_time 1431.14, mean_step_time 1437.9, it/s 0.7
  3525. [0130 17:46:31 @multigpu.py:323] [p0576]  step: count(696), step_time 1501.96, mean_step_time 1441.63, it/s 0.69
  3526. [0130 17:46:32 @multigpu.py:323] [p0576]  step: count(697), step_time 1399.25, mean_step_time 1440.66, it/s 0.69
  3527. [0130 17:46:32 @multigpu.py:323] [p0115]  step: count(674), step_time 1447.63, mean_step_time 1437.0, it/s 0.7
  3528. [0130 17:46:32 @multigpu.py:323] [p0574]  step: count(690), step_time 1477.92, mean_step_time 1447.05, it/s 0.69
  3529. [0130 17:46:34 @multigpu.py:323] [p0574]  step: count(691), step_time 1414.94, mean_step_time 1448.39, it/s 0.69
  3530. [0130 17:46:34 @multigpu.py:323] [p0115]  step: count(675), step_time 1448.86, mean_step_time 1437.88, it/s 0.7
  3531. [0130 17:46:34 @multigpu.py:323] [p0576]  step: count(698), step_time 1515.76, mean_step_time 1447.21, it/s 0.69
  3532. [0130 17:46:35 @multigpu.py:323] [p0115]  step: count(676), step_time 1418.69, mean_step_time 1437.93, it/s 0.7
  3533. [0130 17:46:35 @multigpu.py:323] [p0574]  step: count(692), step_time 1453.11, mean_step_time 1450.68, it/s 0.69
  3534. [0130 17:46:35 @multigpu.py:323] [p0576]  step: count(699), step_time 1455.22, mean_step_time 1445.43, it/s 0.69
  3535. [0130 17:46:36 @multigpu.py:323] [p0115]  step: count(677), step_time 1402.6, mean_step_time 1436.55, it/s 0.7
  3536. [0130 17:46:36 @multigpu.py:323] [p0574]  step: count(693), step_time 1420.45, mean_step_time 1450.0, it/s 0.69
  3537. [0130 17:46:37 @multigpu.py:323] [p0576]  step: count(700), step_time 1405.56, mean_step_time 1446.56, it/s 0.69
  3538. sending debugging info...
  3539. sending to address tcp://p0112:61216
  3540. ##### Sending to neptune:  mean_delay :  0.276642559436 , 0.0 #####
  3541. sending to address tcp://p0112:61216
  3542. ##### Sending to neptune:  max_delay :  0.276642559436 , -0.0 #####
  3543. ##### Sending to neptune:  min_delay :  0.276642559436 , -0.0 #####
  3544. [u'delays', [0.0, -0.0, -0.0]]
  3545. receiving
  3546. ##### Sending to neptune:  cost :  0.276643062168 , -0.00782114826143 #####
  3547. sending to address tcp://p0112:61216
  3548. ##### Sending to neptune:  policy_loss :  0.276643062168 , -0.00080709339818 #####
  3549. ##### Sending to neptune:  xentropy_loss :  0.276643062168 , -2.29106926918 #####
  3550. ##### Sending to neptune:  value_loss :  0.276643062168 , 1.2907692194 #####
  3551. ##### Sending to neptune:  advantage :  0.276643062168 , -9.11761017051e-05 #####
  3552. ##### Sending to neptune:  pred_reward :  0.276643062168 , 0.41138869524 #####
  3553. ##### Sending to neptune:  max_logit :  0.276643062168 , 0.188500359654 #####
  3554. [u'loss', -0.00782114826142788, -0.0008070933981798589, -2.291069269180298, 1.2907692193984985, -9.117610170505941e-05, 0.41138869524002075, 0.18850035965442657]
  3555. receiving
  3556. ##### Sending to neptune:  active_relus :  0.276643491056 , 8843301.94 #####
  3557. ##### Sending to neptune:  dp_per_s :  0.276643491056 , 88.940578809 #####
  3558. [u'other', 8843301.94, 88.94057880896713]
  3559. receiving
  3560. [0130 17:46:38 @multigpu.py:323] [p0574]  step: count(694), step_time 1396.48, mean_step_time 1448.12, it/s 0.69
  3561. [0130 17:46:38 @multigpu.py:323] [p0576]  step: count(701), step_time 1402.01, mean_step_time 1445.33, it/s 0.69
  3562. [0130 17:46:38 @multigpu.py:323] [p0115]  step: count(678), step_time 1503.46, mean_step_time 1436.15, it/s 0.7
  3563. sending to address tcp://p0112:61216
  3564. ##### Sending to neptune:  online_score :  0.277044685218 , 1.8 #####
  3565. [u'online', 1.8]
  3566. receiving
  3567. [0130 17:46:39 @multigpu.py:323] [p0574]  step: count(695), step_time 1431.64, mean_step_time 1449.24, it/s 0.69
  3568. [0130 17:46:39 @multigpu.py:323] [p0115]  step: count(679), step_time 1387.38, mean_step_time 1433.63, it/s 0.7
  3569. [0130 17:46:39 @multigpu.py:323] [p0576]  step: count(702), step_time 1469.67, mean_step_time 1449.9, it/s 0.69
  3570. sending to address tcp://p0112:61216
  3571. ##### Sending to neptune:  online_score :  0.277471321887 , 1.1 #####
  3572. [u'online', 1.1]
  3573. receiving
  3574. [0130 17:46:41 @multigpu.py:323] [p0574]  step: count(696), step_time 1467.27, mean_step_time 1453.2, it/s 0.69
  3575. [0130 17:46:41 @multigpu.py:323] [p0115]  step: count(680), step_time 1459.1, mean_step_time 1432.85, it/s 0.7
  3576. [0130 17:46:41 @multigpu.py:323] [p0576]  step: count(703), step_time 1390.85, mean_step_time 1448.61, it/s 0.69
  3577. sending to address tcp://p0112:61216
  3578. ##### Sending to neptune:  online_score :  0.277992236084 , 1.5 #####
  3579. [u'online', 1.5]
  3580. receiving
  3581. [0130 17:46:42 @multigpu.py:323] [p0574]  step: count(697), step_time 1407.87, mean_step_time 1451.46, it/s 0.69
  3582. [0130 17:46:42 @multigpu.py:323] [p0115]  step: count(681), step_time 1395.23, mean_step_time 1431.79, it/s 0.7
  3583. [0130 17:46:42 @multigpu.py:323] [p0576]  step: count(704), step_time 1424.77, mean_step_time 1449.44, it/s 0.69
  3584. [0130 17:46:44 @multigpu.py:323] [p0115]  step: count(682), step_time 1418.87, mean_step_time 1429.96, it/s 0.7
  3585. [0130 17:46:44 @multigpu.py:323] [p0576]  step: count(705), step_time 1390.71, mean_step_time 1449.1, it/s 0.69
  3586. [0130 17:46:44 @multigpu.py:323] [p0574]  step: count(698), step_time 1461.76, mean_step_time 1454.02, it/s 0.69
  3587. [0130 17:46:45 @multigpu.py:323] [p0115]  step: count(683), step_time 1398.62, mean_step_time 1429.9, it/s 0.7
  3588. [0130 17:46:45 @multigpu.py:323] [p0576]  step: count(706), step_time 1449.88, mean_step_time 1450.04, it/s 0.69
  3589. [0130 17:46:45 @multigpu.py:323] [p0574]  step: count(699), step_time 1457.84, mean_step_time 1456.08, it/s 0.69
  3590. [0130 17:46:46 @multigpu.py:323] [p0576]  step: count(707), step_time 1370.72, mean_step_time 1445.21, it/s 0.69
  3591. [0130 17:46:46 @multigpu.py:323] [p0115]  step: count(684), step_time 1500.55, mean_step_time 1432.84, it/s 0.7
  3592. [0130 17:46:47 @multigpu.py:323] [p0574]  step: count(700), step_time 1450.23, mean_step_time 1457.66, it/s 0.69
  3593. sending debugging info...
  3594. sending to address tcp://p0112:61216
  3595. ##### Sending to neptune:  mean_delay :  0.279424770806 , 0.0 #####
  3596. ##### Sending to neptune:  max_delay :  0.279424770806 , -0.0 #####
  3597. sending to address tcp://p0112:61216
  3598. ##### Sending to neptune:  min_delay :  0.279424770806 , -0.0 #####
  3599. [u'delays', [0.0, -0.0, -0.0]]
  3600. receiving
  3601. ##### Sending to neptune:  cost :  0.279425232477 , -0.00309272995219 #####
  3602. ##### Sending to neptune:  policy_loss :  0.279425232477 , 0.459123522043 #####
  3603. sending to address tcp://p0112:61216
  3604. ##### Sending to neptune:  xentropy_loss :  0.279425232477 , -2.29101920128 #####
  3605. ##### Sending to neptune:  value_loss :  0.279425232477 , 1.43602633476 #####
  3606. ##### Sending to neptune:  advantage :  0.279425232477 , -0.00201497459784 #####
  3607. ##### Sending to neptune:  pred_reward :  0.279425232477 , 0.410510092974 #####
  3608. ##### Sending to neptune:  max_logit :  0.279425232477 , 0.18886230886 #####
  3609. [u'loss', -0.003092729952186346, 0.45912352204322815, -2.2910192012786865, 1.4360263347625732, -0.0020149745978415012, 0.4105100929737091, 0.18886230885982513]
  3610. receiving
  3611. ##### Sending to neptune:  active_relus :  0.279425686664 , 8863210.75 #####
  3612. ##### Sending to neptune:  dp_per_s :  0.279425686664 , 89.0163982094 #####
  3613. [u'other', 8863210.75, 89.01639820936614]
  3614. receiving
  3615. [0130 17:46:48 @multigpu.py:323] [p0576]  step: count(708), step_time 1449.44, mean_step_time 1447.07, it/s 0.69
  3616. [0130 17:46:48 @multigpu.py:323] [p0115]  step: count(685), step_time 1449.69, mean_step_time 1432.79, it/s 0.7
  3617. [0130 17:46:48 @multigpu.py:323] [p0574]  step: count(701), step_time 1451.4, mean_step_time 1458.52, it/s 0.69
  3618. [0130 17:46:49 @multigpu.py:323] [p0115]  step: count(686), step_time 1420.24, mean_step_time 1433.68, it/s 0.7
  3619. [0130 17:46:49 @multigpu.py:323] [p0576]  step: count(709), step_time 1505.57, mean_step_time 1451.1, it/s 0.69
  3620. [0130 17:46:49 @multigpu.py:323] [p0574]  step: count(702), step_time 1430.73, mean_step_time 1461.49, it/s 0.68
  3621. [0130 17:46:51 @multigpu.py:323] [p0576]  step: count(710), step_time 1397.3, mean_step_time 1450.08, it/s 0.69
  3622. [0130 17:46:51 @multigpu.py:323] [p0115]  step: count(687), step_time 1450.65, mean_step_time 1434.68, it/s 0.7
  3623. [0130 17:46:51 @multigpu.py:323] [p0574]  step: count(703), step_time 1462.6, mean_step_time 1461.24, it/s 0.68
  3624. [0130 17:46:52 @multigpu.py:323] [p0115]  step: count(688), step_time 1520.92, mean_step_time 1439.98, it/s 0.69
  3625. [0130 17:46:52 @multigpu.py:323] [p0576]  step: count(711), step_time 1559.96, mean_step_time 1440.03, it/s 0.69
  3626. [0130 17:46:52 @multigpu.py:323] [p0574]  step: count(704), step_time 1453.72, mean_step_time 1440.08, it/s 0.69
  3627. [0130 17:46:54 @multigpu.py:323] [p0574]  step: count(705), step_time 1416.9, mean_step_time 1439.24, it/s 0.69
  3628. [0130 17:46:54 @multigpu.py:323] [p0576]  step: count(712), step_time 1449.48, mean_step_time 1441.87, it/s 0.69
  3629. [0130 17:46:54 @multigpu.py:323] [p0115]  step: count(689), step_time 1481.8, mean_step_time 1441.96, it/s 0.69
  3630. [0130 17:46:55 @multigpu.py:323] [p0574]  step: count(706), step_time 1416.82, mean_step_time 1439.21, it/s 0.69
  3631. [0130 17:46:55 @multigpu.py:323] [p0576]  step: count(713), step_time 1436.16, mean_step_time 1443.14, it/s 0.69
  3632. [0130 17:46:55 @multigpu.py:323] [p0115]  step: count(690), step_time 1442.39, mean_step_time 1444.44, it/s 0.69
  3633. [0130 17:46:57 @multigpu.py:323] [p0574]  step: count(707), step_time 1404.84, mean_step_time 1438.01, it/s 0.7
  3634. [0130 17:46:57 @multigpu.py:323] [p0576]  step: count(714), step_time 1425.54, mean_step_time 1439.99, it/s 0.69
  3635. [0130 17:46:57 @multigpu.py:323] [p0115]  step: count(691), step_time 1429.53, mean_step_time 1444.81, it/s 0.69
  3636. [0130 17:46:58 @multigpu.py:323] [p0576]  step: count(715), step_time 1397.21, mean_step_time 1439.85, it/s 0.69
  3637. [0130 17:46:58 @multigpu.py:323] [p0574]  step: count(708), step_time 1485.55, mean_step_time 1438.76, it/s 0.7
  3638. [0130 17:46:58 @multigpu.py:323] [p0115]  step: count(692), step_time 1433.3, mean_step_time 1442.03, it/s 0.69
  3639. [0130 17:46:59 @multigpu.py:323] [p0576]  step: count(716), step_time 1443.1, mean_step_time 1436.91, it/s 0.7
  3640. [0130 17:47:00 @multigpu.py:323] [p0574]  step: count(709), step_time 1439.46, mean_step_time 1440.08, it/s 0.69
  3641. [0130 17:47:00 @multigpu.py:323] [p0115]  step: count(693), step_time 1417.74, mean_step_time 1441.36, it/s 0.69
  3642. [0130 17:47:01 @multigpu.py:323] [p0574]  step: count(710), step_time 1380.25, mean_step_time 1435.19, it/s 0.7
  3643. [0130 17:47:01 @multigpu.py:323] [p0576]  step: count(717), step_time 1399.54, mean_step_time 1436.92, it/s 0.7
  3644. [0130 17:47:01 @multigpu.py:323] [p0115]  step: count(694), step_time 1413.25, mean_step_time 1439.64, it/s 0.69
  3645. [0130 17:47:02 @multigpu.py:323] [p0576]  step: count(718), step_time 1416.18, mean_step_time 1431.94, it/s 0.7
  3646. [0130 17:47:02 @multigpu.py:323] [p0115]  step: count(695), step_time 1381.65, mean_step_time 1436.28, it/s 0.7
  3647. [0130 17:47:02 @multigpu.py:323] [p0574]  step: count(711), step_time 1458.54, mean_step_time 1437.37, it/s 0.7
  3648. [0130 17:47:04 @multigpu.py:323] [p0576]  step: count(719), step_time 1403.66, mean_step_time 1429.37, it/s 0.7
  3649. [0130 17:47:04 @multigpu.py:323] [p0115]  step: count(696), step_time 1425.07, mean_step_time 1436.6, it/s 0.7
  3650. [0130 17:47:04 @multigpu.py:323] [p0574]  step: count(712), step_time 1508.01, mean_step_time 1440.12, it/s 0.69
  3651. sending to address tcp://p0112:61216
  3652. ##### Sending to neptune:  online_score :  0.284432896111 , 2.0 #####
  3653. [u'online', 2.0]
  3654. receiving
  3655. [0130 17:47:05 @multigpu.py:323] [p0576]  step: count(720), step_time 1396.07, mean_step_time 1428.89, it/s 0.7
  3656. [0130 17:47:05 @multigpu.py:323] [p0115]  step: count(697), step_time 1415.77, mean_step_time 1437.26, it/s 0.7
  3657. [0130 17:47:05 @multigpu.py:323] [p0574]  step: count(713), step_time 1400.91, mean_step_time 1439.14, it/s 0.69
  3658. sending to address tcp://p0112:61216
  3659. ##### Sending to neptune:  online_score :  0.284738871919 , 1.2 #####
  3660. [u'online', 1.2]
  3661. receiving
  3662. [0130 17:47:07 @multigpu.py:323] [p0576]  step: count(721), step_time 1397.28, mean_step_time 1428.65, it/s 0.7
  3663. [0130 17:47:07 @multigpu.py:323] [p0115]  step: count(698), step_time 1417.27, mean_step_time 1432.95, it/s 0.7
  3664. [0130 17:47:07 @multigpu.py:323] [p0574]  step: count(714), step_time 1433.64, mean_step_time 1441.0, it/s 0.69
  3665. [0130 17:47:08 @multigpu.py:323] [p0576]  step: count(722), step_time 1401.82, mean_step_time 1425.26, it/s 0.7
  3666. [0130 17:47:08 @multigpu.py:323] [p0115]  step: count(699), step_time 1443.38, mean_step_time 1435.75, it/s 0.7
  3667. [0130 17:47:08 @multigpu.py:323] [p0574]  step: count(715), step_time 1549.43, mean_step_time 1446.89, it/s 0.69
  3668. sending to address tcp://p0112:61216
  3669. ##### Sending to neptune:  online_score :  0.285529997481 , 1.3 #####
  3670. [u'online', 1.3]
  3671. receiving
  3672. [0130 17:47:09 @multigpu.py:323] [p0576]  step: count(723), step_time 1446.5, mean_step_time 1428.04, it/s 0.7
  3673. [0130 17:47:09 @multigpu.py:323] [p0115]  step: count(700), step_time 1448.74, mean_step_time 1435.23, it/s 0.7
  3674. sending debugging info...
  3675. sending to address tcp://p0112:61216
  3676. ##### Sending to neptune:  mean_delay :  0.285801506374 , 0.0 #####
  3677. sending to address tcp://p0112:61216
  3678. ##### Sending to neptune:  max_delay :  0.285801506374 , -0.0 #####
  3679. ##### Sending to neptune:  min_delay :  0.285801506374 , -0.0 #####
  3680. [u'delays', [0.0, -0.0, -0.0]]
  3681. receiving
  3682. ##### Sending to neptune:  cost :  0.285802074141 , -0.0108789335936 #####
  3683. sending to address tcp://p0112:61216
  3684. ##### Sending to neptune:  policy_loss :  0.285802074141 , -0.324730724096 #####
  3685. ##### Sending to neptune:  xentropy_loss :  0.285802074141 , -2.29079914093 #####
  3686. ##### Sending to neptune:  value_loss :  0.285802074141 , 1.22302639484 #####
  3687. ##### Sending to neptune:  advantage :  0.285802074141 , 0.0014331714483 #####
  3688. ##### Sending to neptune:  pred_reward :  0.285802074141 , 0.409147292376 #####
  3689. ##### Sending to neptune:  max_logit :  0.285802074141 , 0.190351292491 #####
  3690. [u'loss', -0.01087893359363079, -0.3247307240962982, -2.290799140930176, 1.2230263948440552, 0.0014331714482977986, 0.4091472923755646, 0.19035129249095917]
  3691. receiving
  3692. ##### Sending to neptune:  active_relus :  0.285802551111 , 8855540.19 #####
  3693. ##### Sending to neptune:  dp_per_s :  0.285802551111 , 88.9247276744 #####
  3694. [u'other', 8855540.19, 88.92472767440891]
  3695. receiving
  3696. [0130 17:47:10 @multigpu.py:323] [p0574]  step: count(716), step_time 1499.9, mean_step_time 1448.52, it/s 0.69
  3697. [0130 17:47:11 @multigpu.py:323] [p0576]  step: count(724), step_time 1457.24, mean_step_time 1429.67, it/s 0.7
  3698. [0130 17:47:11 @multigpu.py:323] [p0115]  step: count(701), step_time 1416.31, mean_step_time 1436.29, it/s 0.7
  3699. [0130 17:47:11 @multigpu.py:323] [p0574]  step: count(717), step_time 1476.77, mean_step_time 1451.97, it/s 0.69
  3700. [0130 17:47:12 @multigpu.py:323] [p0576]  step: count(725), step_time 1449.38, mean_step_time 1432.6, it/s 0.7
  3701. [0130 17:47:12 @multigpu.py:323] [p0115]  step: count(702), step_time 1464.9, mean_step_time 1438.59, it/s 0.7
  3702. [0130 17:47:13 @multigpu.py:323] [p0574]  step: count(718), step_time 1487.39, mean_step_time 1453.25, it/s 0.69
  3703. [0130 17:47:14 @multigpu.py:323] [p0576]  step: count(726), step_time 1405.97, mean_step_time 1430.41, it/s 0.7
  3704. [0130 17:47:14 @multigpu.py:323] [p0115]  step: count(703), step_time 1397.31, mean_step_time 1438.52, it/s 0.7
  3705. [0130 17:47:14 @multigpu.py:323] [p0574]  step: count(719), step_time 1456.28, mean_step_time 1453.17, it/s 0.69
  3706. [0130 17:47:15 @multigpu.py:323] [p0576]  step: count(727), step_time 1413.06, mean_step_time 1432.52, it/s 0.7
  3707. [0130 17:47:15 @multigpu.py:323] [p0115]  step: count(704), step_time 1476.83, mean_step_time 1437.34, it/s 0.7
  3708. [0130 17:47:16 @multigpu.py:323] [p0574]  step: count(720), step_time 1470.73, mean_step_time 1454.19, it/s 0.69
  3709. [0130 17:47:17 @multigpu.py:323] [p0576]  step: count(728), step_time 1434.16, mean_step_time 1431.76, it/s 0.7
  3710. [0130 17:47:17 @multigpu.py:323] [p0115]  step: count(705), step_time 1408.52, mean_step_time 1435.28, it/s 0.7
  3711. [0130 17:47:17 @multigpu.py:323] [p0574]  step: count(721), step_time 1504.53, mean_step_time 1456.85, it/s 0.69
  3712. [0130 17:47:18 @multigpu.py:323] [p0576]  step: count(729), step_time 1437.14, mean_step_time 1428.34, it/s 0.7
  3713. [0130 17:47:18 @multigpu.py:323] [p0115]  step: count(706), step_time 1473.58, mean_step_time 1437.95, it/s 0.7
  3714. [0130 17:47:19 @multigpu.py:323] [p0574]  step: count(722), step_time 1579.61, mean_step_time 1464.29, it/s 0.68
  3715. [0130 17:47:19 @multigpu.py:323] [p0576]  step: count(730), step_time 1430.95, mean_step_time 1430.02, it/s 0.7
  3716. [0130 17:47:20 @multigpu.py:323] [p0115]  step: count(707), step_time 1453.48, mean_step_time 1438.09, it/s 0.7
  3717. [0130 17:47:20 @multigpu.py:323] [p0574]  step: count(723), step_time 1468.18, mean_step_time 1464.57, it/s 0.68
  3718. [0130 17:47:22 @multigpu.py:323] [p0574]  step: count(724), step_time 1484.95, mean_step_time 1466.14, it/s 0.68
  3719. [0130 17:47:22 @multigpu.py:323] [p0115]  step: count(708), step_time 2084.26, mean_step_time 1466.25, it/s 0.68
  3720. [0130 17:47:22 @multigpu.py:323] [p0576]  step: count(731), step_time 2284.87, mean_step_time 1466.27, it/s 0.68
  3721. [0130 17:47:23 @multigpu.py:323] [p0576]  step: count(732), step_time 1417.49, mean_step_time 1464.67, it/s 0.68
  3722. [0130 17:47:23 @multigpu.py:323] [p0115]  step: count(709), step_time 1423.44, mean_step_time 1463.34, it/s 0.68
  3723. [0130 17:47:23 @multigpu.py:323] [p0574]  step: count(725), step_time 1466.45, mean_step_time 1468.61, it/s 0.68
  3724. [0130 17:47:25 @multigpu.py:323] [p0576]  step: count(733), step_time 1450.55, mean_step_time 1465.39, it/s 0.68
  3725. [0130 17:47:25 @multigpu.py:323] [p0115]  step: count(710), step_time 1444.98, mean_step_time 1463.47, it/s 0.68
  3726. [0130 17:47:25 @multigpu.py:323] [p0574]  step: count(726), step_time 1466.99, mean_step_time 1471.12, it/s 0.68
  3727. [0130 17:47:26 @multigpu.py:323] [p0576]  step: count(734), step_time 1418.3, mean_step_time 1465.02, it/s 0.68
  3728. [0130 17:47:26 @multigpu.py:323] [p0115]  step: count(711), step_time 1459.8, mean_step_time 1464.98, it/s 0.68
  3729. [0130 17:47:26 @multigpu.py:323] [p0574]  step: count(727), step_time 1479.33, mean_step_time 1474.85, it/s 0.68
  3730. [0130 17:47:27 @multigpu.py:323] [p0576]  step: count(735), step_time 1407.67, mean_step_time 1465.55, it/s 0.68
  3731. [0130 17:47:27 @multigpu.py:323] [p0115]  step: count(712), step_time 1419.85, mean_step_time 1464.31, it/s 0.68
  3732. [0130 17:47:28 @multigpu.py:323] [p0574]  step: count(728), step_time 1647.55, mean_step_time 1482.95, it/s 0.67
  3733. [0130 17:47:29 @multigpu.py:323] [p0576]  step: count(736), step_time 1399.97, mean_step_time 1463.39, it/s 0.68
  3734. [0130 17:47:29 @multigpu.py:323] [p0115]  step: count(713), step_time 1468.94, mean_step_time 1466.87, it/s 0.68
  3735. [0130 17:47:29 @multigpu.py:323] [p0574]  step: count(729), step_time 1496.2, mean_step_time 1485.78, it/s 0.67
  3736. [0130 17:47:30 @multigpu.py:323] [p0576]  step: count(737), step_time 1421.17, mean_step_time 1464.47, it/s 0.68
  3737. [0130 17:47:30 @multigpu.py:323] [p0115]  step: count(714), step_time 1443.28, mean_step_time 1468.37, it/s 0.68
  3738. [0130 17:47:31 @multigpu.py:323] [p0574]  step: count(730), step_time 1426.75, mean_step_time 1488.11, it/s 0.67
  3739. [0130 17:47:32 @multigpu.py:323] [p0576]  step: count(738), step_time 1426.87, mean_step_time 1465.01, it/s 0.68
  3740. [0130 17:47:32 @multigpu.py:323] [p0115]  step: count(715), step_time 1442.87, mean_step_time 1471.43, it/s 0.68
  3741. [0130 17:47:32 @multigpu.py:323] [p0574]  step: count(731), step_time 1566.06, mean_step_time 1493.48, it/s 0.67
  3742. [0130 17:47:33 @multigpu.py:323] [p0576]  step: count(739), step_time 1378.06, mean_step_time 1463.73, it/s 0.68
  3743. [0130 17:47:33 @multigpu.py:323] [p0115]  step: count(716), step_time 1397.81, mean_step_time 1470.07, it/s 0.68
  3744. [0130 17:47:34 @multigpu.py:323] [p0574]  step: count(732), step_time 1510.79, mean_step_time 1493.62, it/s 0.67
  3745. [0130 17:47:34 @multigpu.py:323] [p0576]  step: count(740), step_time 1453.38, mean_step_time 1466.59, it/s 0.68
  3746. [0130 17:47:35 @multigpu.py:323] [p0115]  step: count(717), step_time 1452.81, mean_step_time 1471.92, it/s 0.68
  3747. [0130 17:47:35 @multigpu.py:323] [p0574]  step: count(733), step_time 1517.2, mean_step_time 1499.44, it/s 0.67
  3748. [0130 17:47:36 @multigpu.py:323] [p0576]  step: count(741), step_time 1380.13, mean_step_time 1465.73, it/s 0.68
  3749. [0130 17:47:36 @multigpu.py:323] [p0115]  step: count(718), step_time 1455.88, mean_step_time 1473.85, it/s 0.68
  3750. [0130 17:47:37 @multigpu.py:323] [p0574]  step: count(734), step_time 1566.86, mean_step_time 1506.1, it/s 0.66
  3751. [0130 17:47:37 @multigpu.py:323] [p0576]  step: count(742), step_time 1463.09, mean_step_time 1468.8, it/s 0.68
  3752. [0130 17:47:38 @multigpu.py:323] [p0115]  step: count(719), step_time 1447.73, mean_step_time 1474.07, it/s 0.68
  3753. [0130 17:47:38 @multigpu.py:323] [p0574]  step: count(735), step_time 1595.75, mean_step_time 1508.41, it/s 0.66
  3754. [0130 17:47:39 @multigpu.py:323] [p0576]  step: count(743), step_time 1368.8, mean_step_time 1464.91, it/s 0.68
  3755. [0130 17:47:39 @multigpu.py:323] [p0115]  step: count(720), step_time 1455.55, mean_step_time 1474.41, it/s 0.68
  3756. [0130 17:47:40 @multigpu.py:323] [p0574]  step: count(736), step_time 1489.2, mean_step_time 1507.88, it/s 0.66
  3757. [0130 17:47:40 @multigpu.py:323] [p0576]  step: count(744), step_time 1383.44, mean_step_time 1461.22, it/s 0.68
  3758. [0130 17:47:40 @multigpu.py:323] [p0115]  step: count(721), step_time 1451.48, mean_step_time 1476.16, it/s 0.68
  3759. sending to address tcp://p0112:61216
  3760. ##### Sending to neptune:  online_score :  0.294545846913 , 1.8 #####
  3761. [u'online', 1.8]
  3762. ##### Sending to neptune:  active_workers :  0.294546027448 , 3 #####
  3763. receiving
  3764. [0130 17:47:41 @multigpu.py:323] [p0576]  step: count(745), step_time 1408.92, mean_step_time 1459.2, it/s 0.69
  3765. [0130 17:47:42 @multigpu.py:323] [p0574]  step: count(737), step_time 1684.3, mean_step_time 1518.26, it/s 0.66
  3766. [0130 17:47:42 @multigpu.py:323] [p0115]  step: count(722), step_time 1436.37, mean_step_time 1474.74, it/s 0.68
  3767. sending to address tcp://p0112:61216
  3768. ##### Sending to neptune:  online_score :  0.294824038612 , 1.2 #####
  3769. [u'online', 1.2]
  3770. receiving
  3771. sending to address tcp://p0112:61216
  3772. ##### Sending to neptune:  online_score :  0.294907082717 , 2.2 #####
  3773. [u'online', 2.2]
  3774. receiving
  3775. [0130 17:47:43 @multigpu.py:323] [p0576]  step: count(746), step_time 1430.48, mean_step_time 1460.42, it/s 0.68
  3776. [0130 17:47:43 @multigpu.py:323] [p0574]  step: count(738), step_time 1636.61, mean_step_time 1525.72, it/s 0.66
  3777. [0130 17:47:43 @multigpu.py:323] [p0115]  step: count(723), step_time 1376.08, mean_step_time 1473.68, it/s 0.68
  3778. [0130 17:47:44 @multigpu.py:323] [p0576]  step: count(747), step_time 1489.13, mean_step_time 1464.23, it/s 0.68
  3779. [0130 17:47:45 @multigpu.py:323] [p0115]  step: count(724), step_time 1441.64, mean_step_time 1471.92, it/s 0.68
  3780. [0130 17:47:45 @multigpu.py:323] [p0574]  step: count(739), step_time 1543.1, mean_step_time 1530.06, it/s 0.65
  3781. [0130 17:47:46 @multigpu.py:323] [p0576]  step: count(748), step_time 1405.09, mean_step_time 1462.77, it/s 0.68
  3782. [0130 17:47:46 @multigpu.py:323] [p0115]  step: count(725), step_time 1436.45, mean_step_time 1473.31, it/s 0.68
  3783. [0130 17:47:46 @multigpu.py:323] [p0574]  step: count(740), step_time 1569.09, mean_step_time 1534.98, it/s 0.65
  3784. [0130 17:47:47 @multigpu.py:323] [p0576]  step: count(749), step_time 1407.46, mean_step_time 1461.29, it/s 0.68
  3785. [0130 17:47:48 @multigpu.py:323] [p0115]  step: count(726), step_time 1458.22, mean_step_time 1472.55, it/s 0.68
  3786. [0130 17:47:48 @multigpu.py:323] [p0574]  step: count(741), step_time 1437.3, mean_step_time 1531.61, it/s 0.65
  3787. [0130 17:47:49 @multigpu.py:323] [p0576]  step: count(750), step_time 1427.61, mean_step_time 1461.12, it/s 0.68
  3788. [0130 17:47:49 @multigpu.py:323] [p0115]  step: count(727), step_time 1435.61, mean_step_time 1471.65, it/s 0.68
  3789. [0130 17:47:49 @multigpu.py:323] [p0574]  step: count(742), step_time 1468.07, mean_step_time 1526.04, it/s 0.66
  3790. [0130 17:47:50 @multigpu.py:323] [p0576]  step: count(751), step_time 1362.66, mean_step_time 1415.01, it/s 0.71
  3791. [0130 17:47:51 @multigpu.py:323] [p0115]  step: count(728), step_time 2405.93, mean_step_time 1487.74, it/s 0.67
  3792. [0130 17:47:51 @multigpu.py:323] [p0574]  step: count(743), step_time 2188.64, mean_step_time 1562.06, it/s 0.64
  3793. [0130 17:47:51 @multigpu.py:323] [p0576]  step: count(752), step_time 1456.0, mean_step_time 1416.94, it/s 0.71
  3794. [0130 17:47:53 @multigpu.py:323] [p0576]  step: count(753), step_time 1413.45, mean_step_time 1415.08, it/s 0.71
  3795. [0130 17:47:53 @multigpu.py:323] [p0115]  step: count(729), step_time 1422.24, mean_step_time 1487.68, it/s 0.67
  3796. [0130 17:47:53 @multigpu.py:323] [p0574]  step: count(744), step_time 1479.45, mean_step_time 1561.78, it/s 0.64
  3797. [0130 17:47:54 @multigpu.py:323] [p0576]  step: count(754), step_time 1401.71, mean_step_time 1414.25, it/s 0.71
  3798. [0130 17:47:54 @multigpu.py:323] [p0115]  step: count(730), step_time 1433.0, mean_step_time 1487.08, it/s 0.67
  3799. [0130 17:47:54 @multigpu.py:323] [p0574]  step: count(745), step_time 1522.8, mean_step_time 1564.6, it/s 0.64
  3800. [0130 17:47:56 @multigpu.py:323] [p0576]  step: count(755), step_time 1378.85, mean_step_time 1412.81, it/s 0.71
  3801. [0130 17:47:56 @multigpu.py:323] [p0115]  step: count(731), step_time 1391.26, mean_step_time 1483.65, it/s 0.67
  3802. [0130 17:47:56 @multigpu.py:323] [p0574]  step: count(746), step_time 1454.22, mean_step_time 1563.96, it/s 0.64
  3803. [0130 17:47:57 @multigpu.py:323] [p0576]  step: count(756), step_time 1429.84, mean_step_time 1414.31, it/s 0.71
  3804. [0130 17:47:57 @multigpu.py:323] [p0115]  step: count(732), step_time 1433.66, mean_step_time 1484.34, it/s 0.67
  3805. [0130 17:47:57 @multigpu.py:323] [p0574]  step: count(747), step_time 1442.47, mean_step_time 1562.12, it/s 0.64
  3806. [0130 17:47:58 @multigpu.py:323] [p0576]  step: count(757), step_time 1381.52, mean_step_time 1412.32, it/s 0.71
  3807. [0130 17:47:59 @multigpu.py:323] [p0115]  step: count(733), step_time 1428.08, mean_step_time 1482.3, it/s 0.67
  3808. [0130 17:47:59 @multigpu.py:323] [p0574]  step: count(748), step_time 1438.44, mean_step_time 1551.67, it/s 0.64
  3809. [0130 17:48:00 @multigpu.py:323] [p0576]  step: count(758), step_time 1397.49, mean_step_time 1410.85, it/s 0.71
  3810. [0130 17:48:00 @multigpu.py:323] [p0115]  step: count(734), step_time 1458.07, mean_step_time 1483.04, it/s 0.67
  3811. [0130 17:48:00 @multigpu.py:323] [p0574]  step: count(749), step_time 1499.59, mean_step_time 1551.84, it/s 0.64
  3812. [0130 17:48:01 @multigpu.py:323] [p0576]  step: count(759), step_time 1388.58, mean_step_time 1411.38, it/s 0.71
  3813. [0130 17:48:01 @multigpu.py:323] [p0115]  step: count(735), step_time 1409.96, mean_step_time 1481.39, it/s 0.68
  3814. [0130 17:48:02 @multigpu.py:323] [p0574]  step: count(750), step_time 1420.14, mean_step_time 1551.5, it/s 0.64
  3815. [0130 17:48:03 @multigpu.py:323] [p0576]  step: count(760), step_time 1410.87, mean_step_time 1409.26, it/s 0.71
  3816. sending to address tcp://p0112:61216
  3817. ##### Sending to neptune:  online_score :  0.300571438273 , 1.3 #####
  3818. [u'online', 1.3]
  3819. receiving
  3820. [0130 17:48:03 @multigpu.py:323] [p0115]  step: count(736), step_time 1424.39, mean_step_time 1482.72, it/s 0.67
  3821. [0130 17:48:03 @multigpu.py:323] [p0574]  step: count(751), step_time 1604.72, mean_step_time 1553.44, it/s 0.64
  3822. [0130 17:48:04 @multigpu.py:323] [p0576]  step: count(761), step_time 1423.25, mean_step_time 1411.41, it/s 0.71
  3823. [0130 17:48:04 @multigpu.py:323] [p0115]  step: count(737), step_time 1482.03, mean_step_time 1484.18, it/s 0.67
  3824. [0130 17:48:05 @multigpu.py:323] [p0574]  step: count(752), step_time 1496.24, mean_step_time 1552.71, it/s 0.64
  3825. sending to address tcp://p0112:61216
  3826. ##### Sending to neptune:  online_score :  0.301201084124 , 1.1 #####
  3827. [u'online', 1.1]
  3828. receiving
  3829. [0130 17:48:06 @multigpu.py:323] [p0576]  step: count(762), step_time 1436.6, mean_step_time 1410.09, it/s 0.71
  3830. [0130 17:48:06 @multigpu.py:323] [p0115]  step: count(738), step_time 1408.71, mean_step_time 1481.82, it/s 0.67
  3831. [0130 17:48:06 @multigpu.py:323] [p0574]  step: count(753), step_time 1493.95, mean_step_time 1551.55, it/s 0.64
  3832. [0130 17:48:07 @multigpu.py:323] [p0576]  step: count(763), step_time 1437.35, mean_step_time 1413.51, it/s 0.71
  3833. [0130 17:48:07 @multigpu.py:323] [p0115]  step: count(739), step_time 1434.51, mean_step_time 1481.16, it/s 0.68
  3834. [0130 17:48:08 @multigpu.py:323] [p0574]  step: count(754), step_time 1429.3, mean_step_time 1544.67, it/s 0.65
  3835. [0130 17:48:08 @multigpu.py:323] [p0576]  step: count(764), step_time 1373.29, mean_step_time 1413.01, it/s 0.71
  3836. [0130 17:48:09 @multigpu.py:323] [p0115]  step: count(740), step_time 1423.09, mean_step_time 1479.54, it/s 0.68
  3837. [0130 17:48:09 @multigpu.py:323] [p0574]  step: count(755), step_time 1581.44, mean_step_time 1543.95, it/s 0.65
  3838. [0130 17:48:10 @multigpu.py:323] [p0576]  step: count(765), step_time 1404.71, mean_step_time 1412.8, it/s 0.71
  3839. [0130 17:48:10 @multigpu.py:323] [p0115]  step: count(741), step_time 1420.29, mean_step_time 1477.98, it/s 0.68
  3840. [0130 17:48:11 @multigpu.py:323] [p0574]  step: count(756), step_time 1490.79, mean_step_time 1544.03, it/s 0.65
  3841. [0130 17:48:11 @multigpu.py:323] [p0576]  step: count(766), step_time 1401.43, mean_step_time 1411.34, it/s 0.71
  3842. [0130 17:48:12 @multigpu.py:323] [p0115]  step: count(742), step_time 1500.94, mean_step_time 1481.21, it/s 0.68
  3843. [0130 17:48:12 @multigpu.py:323] [p0574]  step: count(757), step_time 1599.73, mean_step_time 1539.8, it/s 0.65
  3844. [0130 17:48:13 @multigpu.py:323] [p0576]  step: count(767), step_time 1473.04, mean_step_time 1410.54, it/s 0.71
  3845. [0130 17:48:13 @multigpu.py:323] [p0115]  step: count(743), step_time 1456.1, mean_step_time 1485.21, it/s 0.67
  3846. [0130 17:48:14 @multigpu.py:323] [p0574]  step: count(758), step_time 1563.87, mean_step_time 1536.17, it/s 0.65
  3847. [0130 17:48:14 @multigpu.py:323] [p0576]  step: count(768), step_time 1401.15, mean_step_time 1410.34, it/s 0.71
  3848. [0130 17:48:14 @multigpu.py:323] [p0115]  step: count(744), step_time 1457.8, mean_step_time 1486.02, it/s 0.67
  3849. [0130 17:48:15 @multigpu.py:323] [p0574]  step: count(759), step_time 1448.84, mean_step_time 1531.46, it/s 0.65
  3850. [0130 17:48:15 @multigpu.py:323] [p0576]  step: count(769), step_time 1433.32, mean_step_time 1411.64, it/s 0.71
  3851. [0130 17:48:16 @multigpu.py:323] [p0115]  step: count(745), step_time 1430.88, mean_step_time 1485.74, it/s 0.67
  3852. sending to address tcp://p0112:61216
  3853. ##### Sending to neptune:  online_score :  0.304254507754 , 1.8 #####
  3854. [u'online', 1.8]
  3855. receiving
  3856. [0130 17:48:17 @multigpu.py:323] [p0574]  step: count(760), step_time 1419.68, mean_step_time 1523.98, it/s 0.66
  3857. [0130 17:48:17 @multigpu.py:323] [p0576]  step: count(770), step_time 1421.96, mean_step_time 1411.35, it/s 0.71
  3858. [0130 17:48:17 @multigpu.py:323] [p0115]  step: count(746), step_time 1416.76, mean_step_time 1483.67, it/s 0.67
  3859. [0130 17:48:18 @multigpu.py:323] [p0574]  step: count(761), step_time 1396.02, mean_step_time 1521.92, it/s 0.66
  3860. [0130 17:48:18 @multigpu.py:323] [p0576]  step: count(771), step_time 1417.25, mean_step_time 1414.08, it/s 0.71
  3861. [0130 17:48:19 @multigpu.py:323] [p0115]  step: count(747), step_time 1478.87, mean_step_time 1485.83, it/s 0.67
  3862. [0130 17:48:20 @multigpu.py:323] [p0574]  step: count(762), step_time 1405.99, mean_step_time 1518.82, it/s 0.66
  3863. [0130 17:48:21 @multigpu.py:323] [p0574]  step: count(763), step_time 1430.9, mean_step_time 1480.93, it/s 0.68
  3864. [0130 17:48:21 @multigpu.py:323] [p0115]  step: count(748), step_time 2311.05, mean_step_time 1481.08, it/s 0.68
  3865. [0130 17:48:21 @multigpu.py:323] [p0576]  step: count(772), step_time 2796.65, mean_step_time 1481.11, it/s 0.68
  3866. [0130 17:48:22 @multigpu.py:323] [p0115]  step: count(749), step_time 1422.93, mean_step_time 1481.12, it/s 0.68
  3867. [0130 17:48:23 @multigpu.py:323] [p0574]  step: count(764), step_time 1436.66, mean_step_time 1478.79, it/s 0.68
  3868. [0130 17:48:23 @multigpu.py:323] [p0576]  step: count(773), step_time 1463.31, mean_step_time 1483.61, it/s 0.67
  3869. [0130 17:48:24 @multigpu.py:323] [p0115]  step: count(750), step_time 1418.68, mean_step_time 1480.4, it/s 0.68
  3870. [0130 17:48:24 @multigpu.py:323] [p0576]  step: count(774), step_time 1410.31, mean_step_time 1484.04, it/s 0.67
  3871. [0130 17:48:24 @multigpu.py:323] [p0574]  step: count(765), step_time 1467.2, mean_step_time 1476.01, it/s 0.68
  3872. [0130 17:48:25 @multigpu.py:323] [p0115]  step: count(751), step_time 1426.06, mean_step_time 1482.14, it/s 0.67
  3873. [0130 17:48:25 @multigpu.py:323] [p0576]  step: count(775), step_time 1424.35, mean_step_time 1486.31, it/s 0.67
  3874. [0130 17:48:25 @multigpu.py:323] [p0574]  step: count(766), step_time 1451.39, mean_step_time 1475.87, it/s 0.68
  3875. [0130 17:48:27 @multigpu.py:323] [p0115]  step: count(752), step_time 1385.07, mean_step_time 1479.71, it/s 0.68
  3876. [0130 17:48:27 @multigpu.py:323] [p0576]  step: count(776), step_time 1461.88, mean_step_time 1487.92, it/s 0.67
  3877. [0130 17:48:27 @multigpu.py:323] [p0574]  step: count(767), step_time 1466.48, mean_step_time 1477.07, it/s 0.68
  3878. [0130 17:48:28 @multigpu.py:323] [p0115]  step: count(753), step_time 1429.92, mean_step_time 1479.81, it/s 0.68
  3879. [0130 17:48:28 @multigpu.py:323] [p0576]  step: count(777), step_time 1398.9, mean_step_time 1488.78, it/s 0.67
  3880. [0130 17:48:28 @multigpu.py:323] [p0574]  step: count(768), step_time 1492.78, mean_step_time 1479.79, it/s 0.68
  3881. [0130 17:48:30 @multigpu.py:323] [p0115]  step: count(754), step_time 1413.36, mean_step_time 1477.57, it/s 0.68
  3882. [0130 17:48:30 @multigpu.py:323] [p0576]  step: count(778), step_time 1383.07, mean_step_time 1488.06, it/s 0.67
  3883. [0130 17:48:30 @multigpu.py:323] [p0574]  step: count(769), step_time 1375.72, mean_step_time 1473.59, it/s 0.68
  3884. [0130 17:48:31 @multigpu.py:323] [p0115]  step: count(755), step_time 1414.28, mean_step_time 1477.79, it/s 0.68
  3885. [0130 17:48:31 @multigpu.py:323] [p0576]  step: count(779), step_time 1457.36, mean_step_time 1491.5, it/s 0.67
  3886. [0130 17:48:31 @multigpu.py:323] [p0574]  step: count(770), step_time 1419.43, mean_step_time 1473.56, it/s 0.68
  3887. sending to address tcp://p0112:61216
  3888. ##### Sending to neptune:  online_score :  0.308666802446 , 1.2 #####
  3889. [u'online', 1.2]
  3890. receiving
  3891. [0130 17:48:32 @multigpu.py:323] [p0115]  step: count(756), step_time 1425.67, mean_step_time 1477.85, it/s 0.68
  3892. [0130 17:48:32 @multigpu.py:323] [p0576]  step: count(780), step_time 1396.16, mean_step_time 1490.77, it/s 0.67
  3893. [0130 17:48:33 @multigpu.py:323] [p0574]  step: count(771), step_time 1385.33, mean_step_time 1462.59, it/s 0.68
  3894. [0130 17:48:34 @multigpu.py:323] [p0115]  step: count(757), step_time 1374.16, mean_step_time 1472.46, it/s 0.68
  3895. [0130 17:48:34 @multigpu.py:323] [p0576]  step: count(781), step_time 1418.66, mean_step_time 1490.54, it/s 0.67
  3896. [0130 17:48:34 @multigpu.py:323] [p0574]  step: count(772), step_time 1427.58, mean_step_time 1459.16, it/s 0.69
  3897. [0130 17:48:35 @multigpu.py:323] [p0115]  step: count(758), step_time 1386.69, mean_step_time 1471.36, it/s 0.68
  3898. [0130 17:48:35 @multigpu.py:323] [p0576]  step: count(782), step_time 1421.39, mean_step_time 1489.78, it/s 0.67
  3899. [0130 17:48:35 @multigpu.py:323] [p0574]  step: count(773), step_time 1451.68, mean_step_time 1457.04, it/s 0.69
  3900. [0130 17:48:37 @multigpu.py:323] [p0115]  step: count(759), step_time 1447.37, mean_step_time 1472.0, it/s 0.68
  3901. [0130 17:48:37 @multigpu.py:323] [p0576]  step: count(783), step_time 1463.02, mean_step_time 1491.06, it/s 0.67
  3902. [0130 17:48:37 @multigpu.py:323] [p0574]  step: count(774), step_time 1477.53, mean_step_time 1459.45, it/s 0.69
  3903. [0130 17:48:38 @multigpu.py:323] [p0115]  step: count(760), step_time 1403.82, mean_step_time 1471.04, it/s 0.68
  3904. [0130 17:48:38 @multigpu.py:323] [p0576]  step: count(784), step_time 1419.34, mean_step_time 1493.36, it/s 0.67
  3905. sending to address tcp://p0112:61216
  3906. ##### Sending to neptune:  online_score :  0.310479466054 , 1.7 #####
  3907. [u'online', 1.7]
  3908. receiving
  3909. [0130 17:48:38 @multigpu.py:323] [p0574]  step: count(775), step_time 1417.12, mean_step_time 1451.24, it/s 0.69
  3910. [0130 17:48:39 @multigpu.py:323] [p0115]  step: count(761), step_time 1476.73, mean_step_time 1473.86, it/s 0.68
  3911. [0130 17:48:40 @multigpu.py:323] [p0576]  step: count(785), step_time 1408.38, mean_step_time 1493.55, it/s 0.67
  3912. [0130 17:48:40 @multigpu.py:323] [p0574]  step: count(776), step_time 1408.75, mean_step_time 1447.14, it/s 0.69
  3913. [0130 17:48:41 @multigpu.py:323] [p0115]  step: count(762), step_time 1445.59, mean_step_time 1471.09, it/s 0.68
  3914. [0130 17:48:41 @multigpu.py:323] [p0576]  step: count(786), step_time 1509.02, mean_step_time 1498.93, it/s 0.67
  3915. [0130 17:48:41 @multigpu.py:323] [p0574]  step: count(777), step_time 1415.0, mean_step_time 1437.9, it/s 0.7
  3916. [0130 17:48:42 @multigpu.py:323] [p0115]  step: count(763), step_time 1418.96, mean_step_time 1469.23, it/s 0.68
  3917. [0130 17:48:43 @multigpu.py:323] [p0576]  step: count(787), step_time 1405.37, mean_step_time 1495.54, it/s 0.67
  3918. [0130 17:48:43 @multigpu.py:323] [p0574]  step: count(778), step_time 1384.63, mean_step_time 1428.94, it/s 0.7
  3919. sending to address tcp://p0112:61216
  3920. ##### Sending to neptune:  online_score :  0.311744225224 , 2.4 #####
  3921. [u'online', 2.4]
  3922. ##### Sending to neptune:  active_workers :  0.311744298604 , 3 #####
  3923. receiving
  3924. [0130 17:48:44 @multigpu.py:323] [p0115]  step: count(764), step_time 1405.48, mean_step_time 1466.62, it/s 0.68
  3925. [0130 17:48:44 @multigpu.py:323] [p0576]  step: count(788), step_time 1391.93, mean_step_time 1495.08, it/s 0.67
  3926. [0130 17:48:44 @multigpu.py:323] [p0574]  step: count(779), step_time 1459.28, mean_step_time 1429.46, it/s 0.7
  3927. [0130 17:48:45 @multigpu.py:323] [p0115]  step: count(765), step_time 1454.4, mean_step_time 1467.79, it/s 0.68
  3928. [0130 17:48:45 @multigpu.py:323] [p0576]  step: count(789), step_time 1437.04, mean_step_time 1495.27, it/s 0.67
  3929. [0130 17:48:45 @multigpu.py:323] [p0574]  step: count(780), step_time 1420.63, mean_step_time 1429.51, it/s 0.7
  3930. [0130 17:48:47 @multigpu.py:323] [p0115]  step: count(766), step_time 1402.97, mean_step_time 1467.1, it/s 0.68
  3931. [0130 17:48:47 @multigpu.py:323] [p0576]  step: count(790), step_time 1389.53, mean_step_time 1493.65, it/s 0.67
  3932. [0130 17:48:47 @multigpu.py:323] [p0574]  step: count(781), step_time 1421.5, mean_step_time 1430.78, it/s 0.7
  3933. [0130 17:48:48 @multigpu.py:323] [p0115]  step: count(767), step_time 1469.79, mean_step_time 1466.65, it/s 0.68
  3934. [0130 17:48:48 @multigpu.py:323] [p0576]  step: count(791), step_time 1432.2, mean_step_time 1494.39, it/s 0.67
  3935. [0130 17:48:48 @multigpu.py:323] [p0574]  step: count(782), step_time 1412.91, mean_step_time 1431.13, it/s 0.7
  3936. [0130 17:48:50 @multigpu.py:323] [p0115]  step: count(768), step_time 1627.26, mean_step_time 1432.46, it/s 0.7
  3937. [0130 17:48:50 @multigpu.py:323] [p0576]  step: count(792), step_time 1555.32, mean_step_time 1432.33, it/s 0.7
  3938. [0130 17:48:50 @multigpu.py:323] [p0574]  step: count(783), step_time 1456.03, mean_step_time 1432.38, it/s 0.7
  3939. [0130 17:48:51 @multigpu.py:323] [p0574]  step: count(784), step_time 1414.5, mean_step_time 1431.27, it/s 0.7
  3940. [0130 17:48:51 @multigpu.py:323] [p0115]  step: count(769), step_time 1418.66, mean_step_time 1432.25, it/s 0.7
  3941. [0130 17:48:51 @multigpu.py:323] [p0576]  step: count(793), step_time 1477.83, mean_step_time 1433.05, it/s 0.7
  3942. [0130 17:48:53 @multigpu.py:323] [p0574]  step: count(785), step_time 1395.25, mean_step_time 1427.68, it/s 0.7
  3943. [0130 17:48:53 @multigpu.py:323] [p0576]  step: count(794), step_time 1397.31, mean_step_time 1432.4, it/s 0.7
  3944. [0130 17:48:53 @multigpu.py:323] [p0115]  step: count(770), step_time 1479.11, mean_step_time 1435.27, it/s 0.7
  3945. [0130 17:48:54 @multigpu.py:323] [p0574]  step: count(786), step_time 1441.36, mean_step_time 1427.18, it/s 0.7
  3946. [0130 17:48:54 @multigpu.py:323] [p0576]  step: count(795), step_time 1391.2, mean_step_time 1430.75, it/s 0.7
  3947. [0130 17:48:54 @multigpu.py:323] [p0115]  step: count(771), step_time 1410.43, mean_step_time 1434.49, it/s 0.7
  3948. [0130 17:48:55 @multigpu.py:323] [p0576]  step: count(796), step_time 1407.22, mean_step_time 1428.01, it/s 0.7
  3949. [0130 17:48:55 @multigpu.py:323] [p0115]  step: count(772), step_time 1371.4, mean_step_time 1433.8, it/s 0.7
  3950. [0130 17:48:55 @multigpu.py:323] [p0574]  step: count(787), step_time 1435.45, mean_step_time 1425.62, it/s 0.7
  3951. [0130 17:48:57 @multigpu.py:323] [p0576]  step: count(797), step_time 1437.44, mean_step_time 1429.94, it/s 0.7
  3952. [0130 17:48:57 @multigpu.py:323] [p0115]  step: count(773), step_time 1437.88, mean_step_time 1434.2, it/s 0.7
  3953. [0130 17:48:57 @multigpu.py:323] [p0574]  step: count(788), step_time 1518.88, mean_step_time 1426.93, it/s 0.7
  3954. [0130 17:48:58 @multigpu.py:323] [p0576]  step: count(798), step_time 1373.41, mean_step_time 1429.46, it/s 0.7
  3955. [0130 17:48:58 @multigpu.py:323] [p0115]  step: count(774), step_time 1420.67, mean_step_time 1434.57, it/s 0.7
  3956. [0130 17:48:58 @multigpu.py:323] [p0574]  step: count(789), step_time 1434.05, mean_step_time 1429.85, it/s 0.7
  3957. [0130 17:49:00 @multigpu.py:323] [p0576]  step: count(799), step_time 1400.6, mean_step_time 1426.62, it/s 0.7
  3958. [0130 17:49:00 @multigpu.py:323] [p0115]  step: count(775), step_time 1451.44, mean_step_time 1436.42, it/s 0.7
  3959. sending to address tcp://p0112:61216
  3960. ##### Sending to neptune:  online_score :  0.316423896088 , 1.5 #####
  3961. [u'online', 1.5]
  3962. receiving
  3963. [0130 17:49:00 @multigpu.py:323] [p0574]  step: count(790), step_time 1374.9, mean_step_time 1427.62, it/s 0.7
  3964. [0130 17:49:01 @multigpu.py:323] [p0576]  step: count(800), step_time 1407.41, mean_step_time 1427.18, it/s 0.7
  3965. sending debugging info...
  3966. sending to address tcp://p0112:61216
  3967. ##### Sending to neptune:  mean_delay :  0.316783569985 , 0.0 #####
  3968. ##### Sending to neptune:  max_delay :  0.316783569985 , -0.0 #####
  3969. sending to address tcp://p0112:61216
  3970. ##### Sending to neptune:  min_delay :  0.316783569985 , -0.0 #####
  3971. [u'delays', [0.0, -0.0, -0.0]]
  3972. receiving
  3973. ##### Sending to neptune:  cost :  0.316784079936 , -0.00433429796249 #####
  3974. ##### Sending to neptune:  policy_loss :  0.316784079936 , 0.223585680127 #####
  3975. sending to address tcp://p0112:61216
  3976. ##### Sending to neptune:  xentropy_loss :  0.316784079936 , -2.28890061378 #####
  3977. ##### Sending to neptune:  value_loss :  0.316784079936 , 1.51052486897 #####
  3978. ##### Sending to neptune:  advantage :  0.316784079936 , -0.00105175154749 #####
  3979. ##### Sending to neptune:  pred_reward :  0.316784079936 , 0.432729452848 #####
  3980. ##### Sending to neptune:  max_logit :  0.316784079936 , 0.198514163494 #####
  3981. [u'loss', -0.004334297962486744, 0.22358568012714386, -2.28890061378479, 1.510524868965149, -0.0010517515474930406, 0.43272945284843445, 0.1985141634941101]
  3982. receiving
  3983. ##### Sending to neptune:  active_relus :  0.316784537435 , 9006400.55 #####
  3984. ##### Sending to neptune:  dp_per_s :  0.316784537435 , 88.4437987855 #####
  3985. [u'other', 9006400.55, 88.44379878546097]
  3986. receiving
  3987. [0130 17:49:01 @multigpu.py:323] [p0115]  step: count(776), step_time 1429.33, mean_step_time 1436.61, it/s 0.7
  3988. [0130 17:49:01 @multigpu.py:323] [p0574]  step: count(791), step_time 1407.01, mean_step_time 1428.7, it/s 0.7
  3989. [0130 17:49:02 @multigpu.py:323] [p0576]  step: count(801), step_time 1415.41, mean_step_time 1427.02, it/s 0.7
  3990. [0130 17:49:03 @multigpu.py:323] [p0574]  step: count(792), step_time 1369.7, mean_step_time 1425.81, it/s 0.7
  3991. [0130 17:49:03 @multigpu.py:323] [p0115]  step: count(777), step_time 1390.76, mean_step_time 1437.44, it/s 0.7
  3992. sending to address tcp://p0112:61216
  3993. ##### Sending to neptune:  online_score :  0.317268399994 , 1.7 #####
  3994. [u'online', 1.7]
  3995. receiving
  3996. [0130 17:49:04 @multigpu.py:323] [p0576]  step: count(802), step_time 1496.01, mean_step_time 1430.75, it/s 0.7
  3997. [0130 17:49:04 @multigpu.py:323] [p0574]  step: count(793), step_time 1414.25, mean_step_time 1423.94, it/s 0.7
  3998. [0130 17:49:04 @multigpu.py:323] [p0115]  step: count(778), step_time 1460.37, mean_step_time 1441.12, it/s 0.69
  3999. sending to address tcp://p0112:61216
  4000. ##### Sending to neptune:  online_score :  0.317744605806 , 1.5 #####
  4001. [u'online', 1.5]
  4002. receiving
  4003. [0130 17:49:05 @multigpu.py:323] [p0576]  step: count(803), step_time 1389.31, mean_step_time 1427.06, it/s 0.7
  4004. [0130 17:49:05 @multigpu.py:323] [p0574]  step: count(794), step_time 1399.28, mean_step_time 1420.02, it/s 0.7
  4005. [0130 17:49:05 @multigpu.py:323] [p0115]  step: count(779), step_time 1458.67, mean_step_time 1441.69, it/s 0.69
  4006. [0130 17:49:07 @multigpu.py:323] [p0576]  step: count(804), step_time 1407.45, mean_step_time 1426.47, it/s 0.7
  4007. [0130 17:49:07 @multigpu.py:323] [p0574]  step: count(795), step_time 1521.25, mean_step_time 1425.23, it/s 0.7
  4008. [0130 17:49:07 @multigpu.py:323] [p0115]  step: count(780), step_time 1413.54, mean_step_time 1442.17, it/s 0.69
  4009. [0130 17:49:08 @multigpu.py:323] [p0576]  step: count(805), step_time 1423.43, mean_step_time 1427.22, it/s 0.7
  4010. [0130 17:49:08 @multigpu.py:323] [p0574]  step: count(796), step_time 1395.91, mean_step_time 1424.59, it/s 0.7
  4011. [0130 17:49:08 @multigpu.py:323] [p0115]  step: count(781), step_time 1415.45, mean_step_time 1439.11, it/s 0.69
  4012. [0130 17:49:10 @multigpu.py:323] [p0576]  step: count(806), step_time 1426.55, mean_step_time 1423.1, it/s 0.7
  4013. [0130 17:49:10 @multigpu.py:323] [p0574]  step: count(797), step_time 1403.43, mean_step_time 1424.01, it/s 0.7
  4014. [0130 17:49:10 @multigpu.py:323] [p0115]  step: count(782), step_time 1402.94, mean_step_time 1436.98, it/s 0.7
  4015. [0130 17:49:11 @multigpu.py:323] [p0576]  step: count(807), step_time 1405.85, mean_step_time 1423.12, it/s 0.7
  4016. [0130 17:49:11 @multigpu.py:323] [p0574]  step: count(798), step_time 1417.62, mean_step_time 1425.66, it/s 0.7
  4017. [0130 17:49:11 @multigpu.py:323] [p0115]  step: count(783), step_time 1446.93, mean_step_time 1438.37, it/s 0.7
  4018. [0130 17:49:12 @multigpu.py:323] [p0576]  step: count(808), step_time 1450.79, mean_step_time 1426.07, it/s 0.7
  4019. [0130 17:49:12 @multigpu.py:323] [p0574]  step: count(799), step_time 1408.44, mean_step_time 1423.12, it/s 0.7
  4020. [0130 17:49:13 @multigpu.py:323] [p0115]  step: count(784), step_time 1434.27, mean_step_time 1439.81, it/s 0.69
  4021. [0130 17:49:14 @multigpu.py:323] [p0574]  step: count(800), step_time 1419.21, mean_step_time 1423.05, it/s 0.7
  4022. sending debugging info...
  4023. sending to address tcp://p0112:61216
  4024. ##### Sending to neptune:  mean_delay :  0.32036268466 , 0.0 #####
  4025. ##### Sending to neptune:  max_delay :  0.32036268466 , -0.0 #####
  4026. sending to address tcp://p0112:61216
  4027. ##### Sending to neptune:  min_delay :  0.32036268466 , -0.0 #####
  4028. [u'delays', [0.0, -0.0, -0.0]]
  4029. receiving
  4030. ##### Sending to neptune:  cost :  0.320363213552 , -0.00875152647495 #####
  4031. ##### Sending to neptune:  policy_loss :  0.320363213552 , -0.275862365961 #####
  4032. sending to address tcp://p0112:61216
  4033. ##### Sending to neptune:  xentropy_loss :  0.320363213552 , -2.28863835335 #####
  4034. ##### Sending to neptune:  value_loss :  0.320363213552 , 1.44430506229 #####
  4035. ##### Sending to neptune:  advantage :  0.320363213552 , 0.00120054150466 #####
  4036. ##### Sending to neptune:  pred_reward :  0.320363213552 , 0.438394933939 #####
  4037. ##### Sending to neptune:  max_logit :  0.320363213552 , 0.199205473065 #####
  4038. [u'loss', -0.008751526474952698, -0.27586236596107483, -2.2886383533477783, 1.4443050622940063, 0.0012005415046587586, 0.4383949339389801, 0.19920547306537628]
  4039. [0130 17:49:14 @multigpu.py:323] [p0576]  step: count(809), step_time 1471.75, mean_step_time 1427.8, it/s 0.7
  4040. receiving
  4041. ##### Sending to neptune:  active_relus :  0.320363672442 , 9008079.44 #####
  4042. ##### Sending to neptune:  dp_per_s :  0.320363672442 , 86.8237239399 #####
  4043. [u'other', 9008079.44, 86.82372393990129]
  4044. receiving
  4045. [0130 17:49:14 @multigpu.py:323] [p0115]  step: count(785), step_time 1388.29, mean_step_time 1436.51, it/s 0.7
  4046. [0130 17:49:15 @multigpu.py:323] [p0576]  step: count(810), step_time 1411.18, mean_step_time 1428.88, it/s 0.7
  4047. [0130 17:49:15 @multigpu.py:323] [p0574]  step: count(801), step_time 1438.67, mean_step_time 1423.91, it/s 0.7
  4048. [0130 17:49:15 @multigpu.py:323] [p0115]  step: count(786), step_time 1420.86, mean_step_time 1437.4, it/s 0.7
  4049. [0130 17:49:17 @multigpu.py:323] [p0576]  step: count(811), step_time 1454.29, mean_step_time 1429.99, it/s 0.7
  4050. [0130 17:49:17 @multigpu.py:323] [p0115]  step: count(787), step_time 1397.73, mean_step_time 1433.8, it/s 0.7
  4051. [0130 17:49:17 @multigpu.py:323] [p0574]  step: count(802), step_time 1481.59, mean_step_time 1427.34, it/s 0.7
  4052. [0130 17:49:18 @multigpu.py:323] [p0574]  step: count(803), step_time 1418.38, mean_step_time 1425.46, it/s 0.7
  4053. [0130 17:49:18 @multigpu.py:323] [p0576]  step: count(812), step_time 1467.82, mean_step_time 1425.61, it/s 0.7
  4054. [0130 17:49:18 @multigpu.py:323] [p0115]  step: count(788), step_time 1465.4, mean_step_time 1425.71, it/s 0.7
  4055. [0130 17:49:20 @multigpu.py:323] [p0574]  step: count(804), step_time 1392.49, mean_step_time 1424.36, it/s 0.7
  4056. [0130 17:49:20 @multigpu.py:323] [p0115]  step: count(789), step_time 1419.76, mean_step_time 1425.76, it/s 0.7
  4057. [0130 17:49:20 @multigpu.py:323] [p0576]  step: count(813), step_time 1429.68, mean_step_time 1423.21, it/s 0.7
  4058. [0130 17:49:21 @multigpu.py:323] [p0576]  step: count(814), step_time 1372.1, mean_step_time 1421.95, it/s 0.7
  4059. [0130 17:49:21 @multigpu.py:323] [p0574]  step: count(805), step_time 1411.0, mean_step_time 1425.14, it/s 0.7
  4060. [0130 17:49:21 @multigpu.py:323] [p0115]  step: count(790), step_time 1485.53, mean_step_time 1426.08, it/s 0.7
  4061. [0130 17:49:22 @multigpu.py:323] [p0576]  step: count(815), step_time 1413.62, mean_step_time 1423.07, it/s 0.7
  4062. [0130 17:49:23 @multigpu.py:323] [p0574]  step: count(806), step_time 1455.09, mean_step_time 1425.83, it/s 0.7
  4063. [0130 17:49:23 @multigpu.py:323] [p0115]  step: count(791), step_time 1448.14, mean_step_time 1427.97, it/s 0.7
  4064. sending to address tcp://p0112:61216
  4065. ##### Sending to neptune:  online_score :  0.322972231905 , 1.4 #####
  4066. [u'online', 1.4]
  4067. receiving
  4068. [0130 17:49:24 @multigpu.py:323] [p0576]  step: count(816), step_time 1431.39, mean_step_time 1424.28, it/s 0.7
  4069. [0130 17:49:24 @multigpu.py:323] [p0574]  step: count(807), step_time 1466.36, mean_step_time 1427.37, it/s 0.7
  4070. [0130 17:49:24 @multigpu.py:323] [p0115]  step: count(792), step_time 1426.63, mean_step_time 1430.73, it/s 0.7
  4071. [0130 17:49:25 @multigpu.py:323] [p0576]  step: count(817), step_time 1429.91, mean_step_time 1423.9, it/s 0.7
  4072. [0130 17:49:25 @multigpu.py:323] [p0574]  step: count(808), step_time 1424.38, mean_step_time 1422.65, it/s 0.7
  4073. [0130 17:49:26 @multigpu.py:323] [p0115]  step: count(793), step_time 1473.78, mean_step_time 1432.52, it/s 0.7
  4074. [0130 17:49:27 @multigpu.py:323] [p0576]  step: count(818), step_time 1418.34, mean_step_time 1426.14, it/s 0.7
  4075. [0130 17:49:27 @multigpu.py:323] [p0574]  step: count(809), step_time 1412.76, mean_step_time 1421.59, it/s 0.7
  4076. [0130 17:49:27 @multigpu.py:323] [p0115]  step: count(794), step_time 1480.85, mean_step_time 1435.53, it/s 0.7
  4077. [0130 17:49:28 @multigpu.py:323] [p0576]  step: count(819), step_time 1430.83, mean_step_time 1427.66, it/s 0.7
  4078. [0130 17:49:28 @multigpu.py:323] [p0574]  step: count(810), step_time 1391.33, mean_step_time 1422.41, it/s 0.7
  4079. [0130 17:49:28 @multigpu.py:323] [p0115]  step: count(795), step_time 1477.14, mean_step_time 1436.82, it/s 0.7
  4080. sending to address tcp://p0112:61216
  4081. ##### Sending to neptune:  online_score :  0.324464672738 , 1.3 #####
  4082. [u'online', 1.3]
  4083. receiving
  4084. [0130 17:49:30 @multigpu.py:323] [p0576]  step: count(820), step_time 1396.28, mean_step_time 1427.1, it/s 0.7
  4085. [0130 17:49:30 @multigpu.py:323] [p0574]  step: count(811), step_time 1441.45, mean_step_time 1424.13, it/s 0.7
  4086. [0130 17:49:30 @multigpu.py:323] [p0115]  step: count(796), step_time 1473.22, mean_step_time 1439.01, it/s 0.69
  4087. [0130 17:49:31 @multigpu.py:323] [p0576]  step: count(821), step_time 1408.66, mean_step_time 1426.76, it/s 0.7
  4088. [0130 17:49:31 @multigpu.py:323] [p0574]  step: count(812), step_time 1404.4, mean_step_time 1425.86, it/s 0.7
  4089. [0130 17:49:31 @multigpu.py:323] [p0115]  step: count(797), step_time 1453.41, mean_step_time 1442.15, it/s 0.69
  4090.  80%|#######9  |799/1000[19:[0130 17:49:32 @multigpu.py:323] [p0576]  step: count(822), step_time 1435.22, mean_step_time 1423.72, it/s 0.7
  4091. [0130 17:49:32 @multigpu.py:323] [p0574]  step: count(813), step_time 1444.78, mean_step_time 1427.39, it/s 0.7
  4092. [0130 17:49:33 @multigpu.py:323] [p0115]  step: count(798), step_time 1475.66, mean_step_time 1442.91, it/s 0.69
  4093. [0130 17:49:34 @multigpu.py:323] [p0576]  step: count(823), step_time 1458.8, mean_step_time 1427.2, it/s 0.7
  4094. [0130 17:49:34 @multigpu.py:323] [p0574]  step: count(814), step_time 1431.78, mean_step_time 1429.02, it/s 0.7
  4095. [0130 17:49:34 @multigpu.py:323] [p0115]  step: count(799), step_time 1477.14, mean_step_time 1443.83, it/s 0.69
  4096. [0130 17:49:35 @multigpu.py:323] [p0576]  step: count(824), step_time 1397.17, mean_step_time 1426.68, it/s 0.7
  4097. [0130 17:49:35 @multigpu.py:323] [p0574]  step: count(815), step_time 1377.52, mean_step_time 1421.83, it/s 0.7
  4098. [0130 17:49:36 @multigpu.py:323] [p0115]  step: count(800), step_time 1403.05, mean_step_time 1443.31, it/s 0.69
  4099. sending debugging info...
  4100. sending to address tcp://p0112:61216
  4101. ##### Sending to neptune:  mean_delay :  0.326430359946 , 0.0 #####
  4102. sending to address tcp://p0112:61216
  4103. sending to address tcp://p0112:61216
  4104. [0130 17:49:37 @multigpu.py:323] [p0576]  step: count(825), step_time 1424.09, mean_step_time 1426.72, it/s 0.7
  4105. [0130 17:49:37 @multigpu.py:323] [p0574]  step: count(816), step_time 1485.52, mean_step_time 1426.31, it/s 0.7
  4106. sending to address tcp://p0112:61216
  4107. [0130 17:49:37 @multigpu.py:323] [p0115]  step: count(801), step_time 1467.26, mean_step_time 1445.9, it/s 0.69
  4108. [0130 17:49:38 @multigpu.py:323] [p0576]  step: count(826), step_time 1476.38, mean_step_time 1429.21, it/s 0.7
  4109. [0130 17:49:38 @multigpu.py:323] [p0574]  step: count(817), step_time 1391.94, mean_step_time 1425.74, it/s 0.7
  4110. [0130 17:49:39 @multigpu.py:323] [p0115]  step: count(802), step_time 1377.75, mean_step_time 1444.64, it/s 0.69
  4111. [0130 17:49:40 @multigpu.py:323] [p0576]  step: count(827), step_time 1429.75, mean_step_time 1430.4, it/s 0.7
  4112. [0130 17:49:40 @multigpu.py:323] [p0574]  step: count(818), step_time 1457.73, mean_step_time 1427.74, it/s 0.7
  4113. [0130 17:49:40 @multigpu.py:323] [p0115]  step: count(803), step_time 1449.73, mean_step_time 1444.78, it/s 0.69
  4114. ##### Sending to neptune:  max_delay :  0.326430359946 , -0.0 #####
  4115. ##### Sending to neptune:  min_delay :  0.326430359946 , -0.0 #####
  4116. [u'delays', [0.0, -0.0, -0.0]]
  4117. receiving
  4118. ##### Sending to neptune:  cost :  0.327900048296 , -0.0139869665727 #####
  4119. ##### Sending to neptune:  policy_loss :  0.327900048296 , -0.741453886032 #####
  4120. ##### Sending to neptune:  xentropy_loss :  0.327900048296 , -2.2882630825 #####
  4121. ##### Sending to neptune:  value_loss :  0.327900048296 , 1.23938500881 #####
  4122. ##### Sending to neptune:  advantage :  0.327900048296 , 0.003333655186 #####
  4123. ##### Sending to neptune:  pred_reward :  0.327900048296 , 0.43941155076 #####
  4124. ##### Sending to neptune:  max_logit :  0.327900048296 , 0.20052549243 #####
  4125. [u'loss', -0.013986966572701931, -0.7414538860321045, -2.2882630825042725, 1.2393850088119507, 0.003333655185997486, 0.43941155076026917, 0.20052549242973328]
  4126. receiving
  4127. ##### Sending to neptune:  active_relus :  0.327902424998 , 9061856.05 #####
  4128. ##### Sending to neptune:  dp_per_s :  0.327902424998 , 87.6794406125 #####
  4129. [u'other', 9061856.05, 87.6794406124564]
  4130. receiving
  4131. ##### Sending to neptune:  online_score :  0.327903013031 , 1.2 #####
  4132. [u'online', 1.2]
  4133. receiving
  4134. [0130 17:49:41 @multigpu.py:323] [p0576]  step: count(828), step_time 1446.81, mean_step_time 1430.2, it/s 0.7
  4135. [0130 17:49:41 @multigpu.py:323] [p0574]  step: count(819), step_time 1411.12, mean_step_time 1427.88, it/s 0.7
  4136. [0130 17:49:41 @multigpu.py:323] [p0115]  step: count(804), step_time 1437.33, mean_step_time 1444.93, it/s 0.69
  4137. [0130 17:49:42 @multigpu.py:323] [p0574]  step: count(820), step_time 1378.63, mean_step_time 1425.85, it/s 0.7
  4138. [0130 17:49:43 @multigpu.py:323] [p0576]  step: count(829), step_time 1459.28, mean_step_time 1429.58, it/s 0.7
  4139. [0130 17:49:43 @multigpu.py:323] [p0115]  step: count(805), step_time 1406.53, mean_step_time 1445.84, it/s 0.69
  4140. [0130 17:49:44 @multigpu.py:323] [p0574]  step: count(821), step_time 1396.77, mean_step_time 1423.75, it/s 0.7
  4141. [0130 17:49:44 @multigpu.py:323] [p0576]  step: count(830), step_time 1395.62, mean_step_time 1428.8, it/s 0.7
  4142. [0130 17:49:44 @multigpu.py:323] [p0115]  step: count(806), step_time 1423.04, mean_step_time 1445.95, it/s 0.69
  4143. [0130 17:49:45 @multigpu.py:323] [p0574]  step: count(822), step_time 1447.25, mean_step_time 1422.03, it/s 0.7
  4144. [0130 17:49:45 @multigpu.py:323] [p0576]  step: count(831), step_time 1470.52, mean_step_time 1429.61, it/s 0.7
  4145. [0130 17:49:46 @multigpu.py:323] [p0115]  step: count(807), step_time 1488.77, mean_step_time 1450.51, it/s 0.69
  4146. [0130 17:49:47 @multigpu.py:323] [p0115]  step: count(808), step_time 1465.87, mean_step_time 1450.53, it/s 0.69
  4147. [0130 17:49:47 @multigpu.py:323] [p0576]  step: count(832), step_time 1888.2, mean_step_time 1450.63, it/s 0.69
  4148. [0130 17:49:47 @multigpu.py:323] [p0574]  step: count(823), step_time 1990.13, mean_step_time 1450.62, it/s 0.69
  4149. [0130 17:49:49 @multigpu.py:323] [p0115]  step: count(809), step_time 1385.91, mean_step_time 1448.84, it/s 0.69
  4150. [0130 17:49:49 @multigpu.py:323] [p0576]  step: count(833), step_time 1420.86, mean_step_time 1450.19, it/s 0.69
  4151. [0130 17:49:49 @multigpu.py:323] [p0574]  step: count(824), step_time 1462.15, mean_step_time 1454.1, it/s 0.69
  4152. [0130 17:49:50 @multigpu.py:323] [p0115]  step: count(810), step_time 1418.66, mean_step_time 1445.49, it/s 0.69
  4153. [0130 17:49:50 @multigpu.py:323] [p0576]  step: count(834), step_time 1409.05, mean_step_time 1452.04, it/s 0.69
  4154. [0130 17:49:50 @multigpu.py:323] [p0574]  step: count(825), step_time 1411.84, mean_step_time 1454.15, it/s 0.69
  4155. [0130 17:49:51 @multigpu.py:323] [p0115]  step: count(811), step_time 1412.64, mean_step_time 1443.72, it/s 0.69
  4156. [0130 17:49:52 @multigpu.py:323] [p0576]  step: count(835), step_time 1400.96, mean_step_time 1451.41, it/s 0.69
  4157. [0130 17:49:52 @multigpu.py:323] [p0574]  step: count(826), step_time 1382.74, mean_step_time 1450.53, it/s 0.69
  4158. [0130 17:49:53 @multigpu.py:323] [p0115]  step: count(812), step_time 1422.38, mean_step_time 1443.5, it/s 0.69
  4159. [0130 17:49:53 @multigpu.py:323] [p0574]  step: count(827), step_time 1434.28, mean_step_time 1448.92, it/s 0.69
  4160. [0130 17:49:53 @multigpu.py:323] [p0576]  step: count(836), step_time 1509.3, mean_step_time 1455.3, it/s 0.69
  4161. sending to address tcp://p0112:61216
  4162. ##### Sending to neptune:  online_score :  0.33134215388 , 0.8 #####
  4163. [u'online', 0.8]
  4164. ##### Sending to neptune:  active_workers :  0.331342185802 , 3 #####
  4165. receiving
  4166. [0130 17:49:54 @multigpu.py:323] [p0115]  step: count(813), step_time 1456.38, mean_step_time 1442.63, it/s 0.69
  4167. [0130 17:49:54 @multigpu.py:323] [p0574]  step: count(828), step_time 1431.32, mean_step_time 1449.27, it/s 0.69
  4168. [0130 17:49:54 @multigpu.py:323] [p0576]  step: count(837), step_time 1418.01, mean_step_time 1454.71, it/s 0.69
  4169. [0130 17:49:56 @multigpu.py:323] [p0115]  step: count(814), step_time 1406.15, mean_step_time 1438.9, it/s 0.69
  4170. [0130 17:49:56 @multigpu.py:323] [p0574]  step: count(829), step_time 1429.44, mean_step_time 1450.11, it/s 0.69
  4171. [0130 17:49:56 @multigpu.py:323] [p0576]  step: count(838), step_time 1479.45, mean_step_time 1457.76, it/s 0.69
  4172. sending to address tcp://p0112:61216
  4173. ##### Sending to neptune:  online_score :  0.332217746642 , 0.9 #####
  4174. [u'online', 0.9]
  4175. receiving
  4176. [0130 17:49:57 @multigpu.py:323] [p0115]  step: count(815), step_time 1472.49, mean_step_time 1438.67, it/s 0.7
  4177. [0130 17:49:57 @multigpu.py:323] [p0574]  step: count(830), step_time 1459.66, mean_step_time 1453.52, it/s 0.69
  4178. [0130 17:49:57 @multigpu.py:323] [p0576]  step: count(839), step_time 1454.89, mean_step_time 1458.96, it/s 0.69
  4179. [0130 17:49:59 @multigpu.py:323] [p0115]  step: count(816), step_time 1392.18, mean_step_time 1434.61, it/s 0.7
  4180. [0130 17:49:59 @multigpu.py:323] [p0574]  step: count(831), step_time 1366.21, mean_step_time 1449.76, it/s 0.69
  4181. [0130 17:49:59 @multigpu.py:323] [p0576]  step: count(840), step_time 1417.42, mean_step_time 1460.02, it/s 0.68
  4182. sending to address tcp://p0112:61216
  4183. ##### Sending to neptune:  online_score :  0.332876266903 , 1.4 #####
  4184. [u'online', 1.4]
  4185. receiving
  4186. [0130 17:50:00 @multigpu.py:323] [p0574]  step: count(832), step_time 1406.13, mean_step_time 1449.85, it/s 0.69
  4187. [0130 17:50:00 @multigpu.py:323] [p0115]  step: count(817), step_time 1455.63, mean_step_time 1434.73, it/s 0.7
  4188. [0130 17:50:00 @multigpu.py:323] [p0576]  step: count(841), step_time 1411.56, mean_step_time 1460.17, it/s 0.68
  4189.  81%|########  |806/1000[19:[0130 17:50:02 @multigpu.py:323] [p0574]  step: count(833), step_time 1442.25, mean_step_time 1449.72, it/s 0.69
  4190. [0130 17:50:02 @multigpu.py:323] [p0115]  step: count(818), step_time 1425.83, mean_step_time 1432.23, it/s 0.7
  4191. [0130 17:50:02 @multigpu.py:323] [p0576]  step: count(842), step_time 1418.03, mean_step_time 1459.31, it/s 0.69
  4192. [0130 17:50:03 @multigpu.py:323] [p0574]  step: count(834), step_time 1427.28, mean_step_time 1449.5, it/s 0.69
  4193. [0130 17:50:03 @multigpu.py:323] [p0115]  step: count(819), step_time 1511.54, mean_step_time 1433.95, it/s 0.7
  4194. [0130 17:50:03 @multigpu.py:323] [p0576]  step: count(843), step_time 1472.28, mean_step_time 1459.98, it/s 0.68
  4195. [0130 17:50:04 @multigpu.py:323] [p0574]  step: count(835), step_time 1462.29, mean_step_time 1453.73, it/s 0.69
  4196. [0130 17:50:04 @multigpu.py:323] [p0115]  step: count(820), step_time 1412.58, mean_step_time 1434.43, it/s 0.7
  4197. [0130 17:50:04 @multigpu.py:323] [p0576]  step: count(844), step_time 1389.04, mean_step_time 1459.57, it/s 0.69
  4198. [0130 17:50:06 @multigpu.py:323] [p0574]  step: count(836), step_time 1450.97, mean_step_time 1452.01, it/s 0.69
  4199. [0130 17:50:06 @multigpu.py:323] [p0115]  step: count(821), step_time 1431.96, mean_step_time 1432.67, it/s 0.7
  4200. [0130 17:50:06 @multigpu.py:323] [p0576]  step: count(845), step_time 1426.12, mean_step_time 1459.68, it/s 0.69
  4201. [0130 17:50:07 @multigpu.py:323] [p0574]  step: count(837), step_time 1483.71, mean_step_time 1456.59, it/s 0.69
  4202. [0130 17:50:07 @multigpu.py:323] [p0576]  step: count(846), step_time 1441.06, mean_step_time 1457.91, it/s 0.69
  4203. [0130 17:50:07 @multigpu.py:323] [p0115]  step: count(822), step_time 1465.82, mean_step_time 1437.07, it/s 0.7
  4204. [0130 17:50:09 @multigpu.py:323] [p0574]  step: count(838), step_time 1427.26, mean_step_time 1455.07, it/s 0.69
  4205. [0130 17:50:09 @multigpu.py:323] [p0115]  step: count(823), step_time 1444.49, mean_step_time 1436.81, it/s 0.7
  4206. [0130 17:50:09 @multigpu.py:323] [p0576]  step: count(847), step_time 1471.8, mean_step_time 1460.01, it/s 0.68
  4207. [0130 17:50:10 @multigpu.py:323] [p0574]  step: count(839), step_time 1411.78, mean_step_time 1455.1, it/s 0.69
  4208. [0130 17:50:10 @multigpu.py:323] [p0576]  step: count(848), step_time 1398.02, mean_step_time 1457.57, it/s 0.69
  4209. [0130 17:50:10 @multigpu.py:323] [p0115]  step: count(824), step_time 1426.87, mean_step_time 1436.29, it/s 0.7
  4210. [0130 17:50:12 @multigpu.py:323] [p0574]  step: count(840), step_time 1421.72, mean_step_time 1457.26, it/s 0.69
  4211. [0130 17:50:12 @multigpu.py:323] [p0576]  step: count(849), step_time 1433.66, mean_step_time 1456.29, it/s 0.69
  4212. [0130 17:50:12 @multigpu.py:323] [p0115]  step: count(825), step_time 1437.46, mean_step_time 1437.83, it/s 0.7
  4213. [0130 17:50:13 @multigpu.py:323] [p0574]  step: count(841), step_time 1459.84, mean_step_time 1460.41, it/s 0.68
  4214. [0130 17:50:13 @multigpu.py:323] [p0115]  step: count(826), step_time 1400.72, mean_step_time 1436.72, it/s 0.7
  4215. [0130 17:50:13 @multigpu.py:323] [p0576]  step: count(850), step_time 1425.38, mean_step_time 1457.78, it/s 0.69
  4216. [0130 17:50:14 @multigpu.py:323] [p0115]  step: count(827), step_time 1374.98, mean_step_time 1431.03, it/s 0.7
  4217. [0130 17:50:14 @multigpu.py:323] [p0574]  step: count(842), step_time 1388.76, mean_step_time 1457.49, it/s 0.69
  4218. [0130 17:50:15 @multigpu.py:323] [p0576]  step: count(851), step_time 1446.51, mean_step_time 1456.58, it/s 0.69
  4219. [0130 17:50:16 @multigpu.py:323] [p0115]  step: count(828), step_time 1508.09, mean_step_time 1433.14, it/s 0.7
  4220. [0130 17:50:16 @multigpu.py:323] [p0576]  step: count(852), step_time 1418.63, mean_step_time 1433.1, it/s 0.7
  4221. [0130 17:50:16 @multigpu.py:323] [p0574]  step: count(843), step_time 1501.38, mean_step_time 1433.05, it/s 0.7
  4222. [0130 17:50:17 @multigpu.py:323] [p0576]  step: count(853), step_time 1382.78, mean_step_time 1431.2, it/s 0.7
  4223. [0130 17:50:17 @multigpu.py:323] [p0115]  step: count(829), step_time 1417.21, mean_step_time 1434.7, it/s 0.7
  4224. [0130 17:50:17 @multigpu.py:323] [p0574]  step: count(844), step_time 1492.5, mean_step_time 1434.57, it/s 0.7
  4225. [0130 17:50:19 @multigpu.py:323] [p0576]  step: count(854), step_time 1431.19, mean_step_time 1432.3, it/s 0.7
  4226. [0130 17:50:19 @multigpu.py:323] [p0115]  step: count(830), step_time 1462.53, mean_step_time 1436.9, it/s 0.7
  4227. [0130 17:50:19 @multigpu.py:323] [p0574]  step: count(845), step_time 1454.74, mean_step_time 1436.71, it/s 0.7
  4228. [0130 17:50:20 @multigpu.py:323] [p0576]  step: count(855), step_time 1445.91, mean_step_time 1434.55, it/s 0.7
  4229. [0130 17:50:20 @multigpu.py:323] [p0115]  step: count(831), step_time 1409.05, mean_step_time 1436.72, it/s 0.7
  4230. [0130 17:50:20 @multigpu.py:323] [p0574]  step: count(846), step_time 1417.08, mean_step_time 1438.43, it/s 0.7
  4231. [0130 17:50:22 @multigpu.py:323] [p0576]  step: count(856), step_time 1438.58, mean_step_time 1431.02, it/s 0.7
  4232. [0130 17:50:22 @multigpu.py:323] [p0115]  step: count(832), step_time 1501.15, mean_step_time 1440.66, it/s 0.69
  4233. [0130 17:50:22 @multigpu.py:323] [p0574]  step: count(847), step_time 1478.96, mean_step_time 1440.66, it/s 0.69
  4234. [0130 17:50:23 @multigpu.py:323] [p0576]  step: count(857), step_time 1419.0, mean_step_time 1431.07, it/s 0.7
  4235. [0130 17:50:23 @multigpu.py:323] [p0115]  step: count(833), step_time 1378.58, mean_step_time 1436.77, it/s 0.7
  4236. [0130 17:50:23 @multigpu.py:323] [p0574]  step: count(848), step_time 1425.69, mean_step_time 1440.38, it/s 0.69
  4237. sending to address tcp://p0112:61216
  4238. ##### Sending to neptune:  online_score :  0.339696066909 , 1.6 #####
  4239. [u'online', 1.6]
  4240. receiving
  4241. sending to address tcp://p0112:61216
  4242. ##### Sending to neptune:  online_score :  0.339726577732 , 2.1 #####
  4243. [u'online', 2.1]
  4244. receiving
  4245. sending to address tcp://p0112:61216
  4246. ##### Sending to neptune:  online_score :  0.339784356621 , 1.0 #####
  4247. [u'online', 1.0]
  4248. receiving
  4249. [0130 17:50:24 @multigpu.py:323] [p0576]  step: count(858), step_time 1418.17, mean_step_time 1428.0, it/s 0.7
  4250. [0130 17:50:25 @multigpu.py:323] [p0115]  step: count(834), step_time 1484.33, mean_step_time 1440.68, it/s 0.69
  4251. [0130 17:50:25 @multigpu.py:323] [p0574]  step: count(849), step_time 1422.01, mean_step_time 1440.01, it/s 0.69
  4252. [0130 17:50:26 @multigpu.py:323] [p0576]  step: count(859), step_time 1401.0, mean_step_time 1425.31, it/s 0.7
  4253. [0130 17:50:26 @multigpu.py:323] [p0115]  step: count(835), step_time 1418.18, mean_step_time 1437.96, it/s 0.7
  4254. [0130 17:50:26 @multigpu.py:323] [p0574]  step: count(850), step_time 1462.0, mean_step_time 1440.13, it/s 0.69
  4255. [0130 17:50:27 @multigpu.py:323] [p0576]  step: count(860), step_time 1458.91, mean_step_time 1427.38, it/s 0.7
  4256. [0130 17:50:27 @multigpu.py:323] [p0115]  step: count(836), step_time 1469.8, mean_step_time 1441.84, it/s 0.69
  4257. [0130 17:50:28 @multigpu.py:323] [p0574]  step: count(851), step_time 1412.3, mean_step_time 1442.43, it/s 0.69
  4258. [0130 17:50:29 @multigpu.py:323] [p0576]  step: count(861), step_time 1416.72, mean_step_time 1427.64, it/s 0.7
  4259. [0130 17:50:29 @multigpu.py:323] [p0115]  step: count(837), step_time 1418.69, mean_step_time 1439.99, it/s 0.69
  4260. [0130 17:50:29 @multigpu.py:323] [p0574]  step: count(852), step_time 1398.8, mean_step_time 1442.07, it/s 0.69
  4261. [0130 17:50:30 @multigpu.py:323] [p0576]  step: count(862), step_time 1441.13, mean_step_time 1428.79, it/s 0.7
  4262. [0130 17:50:30 @multigpu.py:323] [p0115]  step: count(838), step_time 1361.38, mean_step_time 1436.77, it/s 0.7
  4263. [0130 17:50:30 @multigpu.py:323] [p0574]  step: count(853), step_time 1438.35, mean_step_time 1441.87, it/s 0.69
  4264. [0130 17:50:32 @multigpu.py:323] [p0576]  step: count(863), step_time 1437.12, mean_step_time 1427.04, it/s 0.7
  4265. [0130 17:50:32 @multigpu.py:323] [p0115]  step: count(839), step_time 1439.47, mean_step_time 1433.17, it/s 0.7
  4266. [0130 17:50:32 @multigpu.py:323] [p0574]  step: count(854), step_time 1416.04, mean_step_time 1441.31, it/s 0.69
  4267. [0130 17:50:33 @multigpu.py:323] [p0576]  step: count(864), step_time 1436.49, mean_step_time 1429.41, it/s 0.7
  4268. [0130 17:50:33 @multigpu.py:323] [p0115]  step: count(840), step_time 1375.77, mean_step_time 1431.33, it/s 0.7
  4269. [0130 17:50:33 @multigpu.py:323] [p0574]  step: count(855), step_time 1452.58, mean_step_time 1440.82, it/s 0.69
  4270. [0130 17:50:35 @multigpu.py:323] [p0115]  step: count(841), step_time 1487.83, mean_step_time 1434.12, it/s 0.7
  4271. [0130 17:50:35 @multigpu.py:323] [p0576]  step: count(865), step_time 1508.24, mean_step_time 1433.52, it/s 0.7
  4272. [0130 17:50:35 @multigpu.py:323] [p0574]  step: count(856), step_time 1409.41, mean_step_time 1438.75, it/s 0.7
  4273. [0130 17:50:36 @multigpu.py:323] [p0576]  step: count(866), step_time 1412.2, mean_step_time 1432.07, it/s 0.7
  4274. [0130 17:50:36 @multigpu.py:323] [p0115]  step: count(842), step_time 1436.41, mean_step_time 1432.65, it/s 0.7
  4275. [0130 17:50:36 @multigpu.py:323] [p0574]  step: count(857), step_time 1446.84, mean_step_time 1436.9, it/s 0.7
  4276. [0130 17:50:37 @multigpu.py:323] [p0576]  step: count(867), step_time 1435.4, mean_step_time 1430.25, it/s 0.7
  4277. [0130 17:50:37 @multigpu.py:323] [p0574]  step: count(858), step_time 1403.72, mean_step_time 1435.72, it/s 0.7
  4278. [0130 17:50:38 @multigpu.py:323] [p0115]  step: count(843), step_time 1490.01, mean_step_time 1434.93, it/s 0.7
  4279. [0130 17:50:39 @multigpu.py:323] [p0115]  step: count(844), step_time 1376.99, mean_step_time 1432.43, it/s 0.7
  4280. [0130 17:50:39 @multigpu.py:323] [p0576]  step: count(868), step_time 1444.38, mean_step_time 1432.57, it/s 0.7
  4281. [0130 17:50:39 @multigpu.py:323] [p0574]  step: count(859), step_time 1435.49, mean_step_time 1436.91, it/s 0.7
  4282. [0130 17:50:40 @multigpu.py:323] [p0115]  step: count(845), step_time 1388.07, mean_step_time 1429.96, it/s 0.7
  4283. [0130 17:50:40 @multigpu.py:323] [p0576]  step: count(869), step_time 1400.91, mean_step_time 1430.93, it/s 0.7
  4284. [0130 17:50:40 @multigpu.py:323] [p0574]  step: count(860), step_time 1422.81, mean_step_time 1436.96, it/s 0.7
  4285. [0130 17:50:42 @multigpu.py:323] [p0115]  step: count(846), step_time 1405.59, mean_step_time 1430.21, it/s 0.7
  4286. [0130 17:50:42 @multigpu.py:323] [p0576]  step: count(870), step_time 1414.07, mean_step_time 1430.37, it/s 0.7
  4287. [0130 17:50:42 @multigpu.py:323] [p0574]  step: count(861), step_time 1389.45, mean_step_time 1433.45, it/s 0.7
  4288. [0130 17:50:43 @multigpu.py:323] [p0576]  step: count(871), step_time 1439.59, mean_step_time 1430.02, it/s 0.7
  4289. [0130 17:50:43 @multigpu.py:323] [p0115]  step: count(847), step_time 1472.43, mean_step_time 1435.08, it/s 0.7
  4290. [0130 17:50:43 @multigpu.py:323] [p0574]  step: count(862), step_time 1486.02, mean_step_time 1438.31, it/s 0.7
  4291. [0130 17:50:45 @multigpu.py:323] [p0574]  step: count(863), step_time 1380.9, mean_step_time 1432.28, it/s 0.7
  4292. [0130 17:50:45 @multigpu.py:323] [p0115]  step: count(848), step_time 1453.38, mean_step_time 1432.34, it/s 0.7
  4293. [0130 17:50:45 @multigpu.py:323] [p0576]  step: count(872), step_time 1464.71, mean_step_time 1432.33, it/s 0.7
  4294. [0130 17:50:46 @multigpu.py:323] [p0115]  step: count(849), step_time 1356.6, mean_step_time 1429.31, it/s 0.7
  4295. [0130 17:50:46 @multigpu.py:323] [p0576]  step: count(873), step_time 1428.41, mean_step_time 1434.61, it/s 0.7
  4296. [0130 17:50:46 @multigpu.py:323] [p0574]  step: count(864), step_time 1449.39, mean_step_time 1430.13, it/s 0.7
  4297. sending to address tcp://p0112:61216
  4298. ##### Sending to neptune:  online_score :  0.346096348829 , 1.0 #####
  4299. [u'online', 1.0]
  4300. receiving
  4301. [0130 17:50:47 @multigpu.py:323] [p0115]  step: count(850), step_time 1441.86, mean_step_time 1428.28, it/s 0.7
  4302. [0130 17:50:47 @multigpu.py:323] [p0576]  step: count(874), step_time 1386.22, mean_step_time 1432.36, it/s 0.7
  4303. [0130 17:50:48 @multigpu.py:323] [p0574]  step: count(865), step_time 1461.56, mean_step_time 1430.47, it/s 0.7
  4304. [0130 17:50:49 @multigpu.py:323] [p0115]  step: count(851), step_time 1407.82, mean_step_time 1428.22, it/s 0.7
  4305. [0130 17:50:49 @multigpu.py:323] [p0576]  step: count(875), step_time 1401.26, mean_step_time 1430.13, it/s 0.7
  4306. [0130 17:50:49 @multigpu.py:323] [p0574]  step: count(866), step_time 1420.17, mean_step_time 1430.62, it/s 0.7
  4307. [0130 17:50:50 @multigpu.py:323] [p0115]  step: count(852), step_time 1406.47, mean_step_time 1423.48, it/s 0.7
  4308. [0130 17:50:50 @multigpu.py:323] [p0576]  step: count(876), step_time 1468.13, mean_step_time 1431.6, it/s 0.7
  4309. [0130 17:50:50 @multigpu.py:323] [p0574]  step: count(867), step_time 1370.78, mean_step_time 1425.22, it/s 0.7
  4310. [0130 17:50:52 @multigpu.py:323] [p0115]  step: count(853), step_time 1394.32, mean_step_time 1424.27, it/s 0.7
  4311. sending to address tcp://p0112:61216
  4312. ##### Sending to neptune:  online_score :  0.347506924139 , 1.4 #####
  4313. [u'online', 1.4]
  4314. receiving
  4315. [0130 17:50:52 @multigpu.py:323] [p0576]  step: count(877), step_time 1458.1, mean_step_time 1433.56, it/s 0.7
  4316. [0130 17:50:52 @multigpu.py:323] [p0574]  step: count(868), step_time 1443.72, mean_step_time 1426.12, it/s 0.7
  4317. [0130 17:50:53 @multigpu.py:323] [p0115]  step: count(854), step_time 1444.81, mean_step_time 1422.29, it/s 0.7
  4318. [0130 17:50:53 @multigpu.py:323] [p0576]  step: count(878), step_time 1429.86, mean_step_time 1434.14, it/s 0.7
  4319. [0130 17:50:53 @multigpu.py:323] [p0574]  step: count(869), step_time 1440.8, mean_step_time 1427.06, it/s 0.7
  4320. sending to address tcp://p0112:61216
  4321. ##### Sending to neptune:  online_score :  0.348158127997 , 1.8 #####
  4322. [u'online', 1.8]
  4323. ##### Sending to neptune:  active_workers :  0.348158168859 , 3 #####
  4324. receiving
  4325. [0130 17:50:55 @multigpu.py:323] [p0115]  step: count(855), step_time 1461.99, mean_step_time 1424.48, it/s 0.7
  4326. [0130 17:50:55 @multigpu.py:323] [p0576]  step: count(879), step_time 1429.79, mean_step_time 1435.58, it/s 0.7
  4327. [0130 17:50:55 @multigpu.py:323] [p0574]  step: count(870), step_time 1415.5, mean_step_time 1424.73, it/s 0.7
  4328. [0130 17:50:56 @multigpu.py:323] [p0115]  step: count(856), step_time 1440.17, mean_step_time 1423.0, it/s 0.7
  4329. [0130 17:50:56 @multigpu.py:323] [p0574]  step: count(871), step_time 1401.46, mean_step_time 1424.19, it/s 0.7
  4330. [0130 17:50:56 @multigpu.py:323] [p0576]  step: count(880), step_time 1490.08, mean_step_time 1437.14, it/s 0.7
  4331. [0130 17:50:57 @multigpu.py:323] [p0115]  step: count(857), step_time 1438.35, mean_step_time 1423.98, it/s 0.7
  4332. [0130 17:50:57 @multigpu.py:323] [p0574]  step: count(872), step_time 1437.87, mean_step_time 1426.14, it/s 0.7
  4333. [0130 17:50:58 @multigpu.py:323] [p0576]  step: count(881), step_time 1417.36, mean_step_time 1437.17, it/s 0.7
  4334. [0130 17:50:59 @multigpu.py:323] [p0115]  step: count(858), step_time 1433.35, mean_step_time 1427.58, it/s 0.7
  4335. [0130 17:50:59 @multigpu.py:323] [p0574]  step: count(873), step_time 1393.58, mean_step_time 1423.9, it/s 0.7
  4336. [0130 17:50:59 @multigpu.py:323] [p0576]  step: count(882), step_time 1393.8, mean_step_time 1434.81, it/s 0.7
  4337. [0130 17:51:00 @multigpu.py:323] [p0115]  step: count(859), step_time 1424.23, mean_step_time 1426.82, it/s 0.7
  4338. [0130 17:51:00 @multigpu.py:323] [p0576]  step: count(883), step_time 1381.47, mean_step_time 1432.02, it/s 0.7
  4339. [0130 17:51:00 @multigpu.py:323] [p0574]  step: count(874), step_time 1502.83, mean_step_time 1428.24, it/s 0.7
  4340. [0130 17:51:02 @multigpu.py:323] [p0115]  step: count(860), step_time 1398.0, mean_step_time 1427.93, it/s 0.7
  4341. [0130 17:51:02 @multigpu.py:323] [p0576]  step: count(884), step_time 1422.88, mean_step_time 1431.34, it/s 0.7
  4342. [0130 17:51:02 @multigpu.py:323] [p0574]  step: count(875), step_time 1428.62, mean_step_time 1427.05, it/s 0.7
  4343. [0130 17:51:03 @multigpu.py:323] [p0115]  step: count(861), step_time 1420.77, mean_step_time 1424.58, it/s 0.7
  4344. [0130 17:51:03 @multigpu.py:323] [p0576]  step: count(885), step_time 1435.11, mean_step_time 1427.69, it/s 0.7
  4345. [0130 17:51:03 @multigpu.py:323] [p0574]  step: count(876), step_time 1461.93, mean_step_time 1429.67, it/s 0.7
  4346. [0130 17:51:05 @multigpu.py:323] [p0115]  step: count(862), step_time 1455.64, mean_step_time 1425.54, it/s 0.7
  4347. [0130 17:51:05 @multigpu.py:323] [p0576]  step: count(886), step_time 1386.82, mean_step_time 1426.42, it/s 0.7
  4348. [0130 17:51:05 @multigpu.py:323] [p0574]  step: count(877), step_time 1405.52, mean_step_time 1427.61, it/s 0.7
  4349. [0130 17:51:06 @multigpu.py:323] [p0115]  step: count(863), step_time 1414.31, mean_step_time 1421.76, it/s 0.7
  4350. [0130 17:51:06 @multigpu.py:323] [p0576]  step: count(887), step_time 1438.14, mean_step_time 1426.56, it/s 0.7
  4351. [0130 17:51:06 @multigpu.py:323] [p0574]  step: count(878), step_time 1390.07, mean_step_time 1426.92, it/s 0.7
  4352. [0130 17:51:07 @multigpu.py:323] [p0115]  step: count(864), step_time 1426.81, mean_step_time 1424.25, it/s 0.7
  4353. [0130 17:51:07 @multigpu.py:323] [p0576]  step: count(888), step_time 1423.35, mean_step_time 1425.5, it/s 0.7
  4354. [0130 17:51:08 @multigpu.py:323] [p0574]  step: count(879), step_time 1486.03, mean_step_time 1429.45, it/s 0.7
  4355. [0130 17:51:09 @multigpu.py:323] [p0115]  step: count(865), step_time 1462.49, mean_step_time 1427.97, it/s 0.7
  4356. [0130 17:51:09 @multigpu.py:323] [p0576]  step: count(889), step_time 1489.7, mean_step_time 1429.94, it/s 0.7
  4357. [0130 17:51:09 @multigpu.py:323] [p0574]  step: count(880), step_time 1528.76, mean_step_time 1434.75, it/s 0.7
  4358. [0130 17:51:10 @multigpu.py:323] [p0115]  step: count(866), step_time 1439.06, mean_step_time 1429.64, it/s 0.7
  4359. [0130 17:51:10 @multigpu.py:323] [p0576]  step: count(890), step_time 1421.23, mean_step_time 1430.3, it/s 0.7
  4360. [0130 17:51:11 @multigpu.py:323] [p0574]  step: count(881), step_time 1589.16, mean_step_time 1444.73, it/s 0.69
  4361. [0130 17:51:12 @multigpu.py:323] [p0115]  step: count(867), step_time 1414.91, mean_step_time 1426.77, it/s 0.7
  4362. [0130 17:51:12 @multigpu.py:323] [p0576]  step: count(891), step_time 1479.83, mean_step_time 1432.31, it/s 0.7
  4363. [0130 17:51:12 @multigpu.py:323] [p0574]  step: count(882), step_time 1398.45, mean_step_time 1440.35, it/s 0.69
  4364. [0130 17:51:14 @multigpu.py:323] [p0574]  step: count(883), step_time 1463.84, mean_step_time 1444.5, it/s 0.69
  4365. [0130 17:51:14 @multigpu.py:323] [p0115]  step: count(868), step_time 1808.96, mean_step_time 1444.55, it/s 0.69
  4366. [0130 17:51:14 @multigpu.py:323] [p0576]  step: count(892), step_time 1699.42, mean_step_time 1444.05, it/s 0.69
  4367. sending to address tcp://p0112:61216
  4368. ##### Sending to neptune:  online_score :  0.353776509696 , 1.7 #####
  4369. [u'online', 1.7]
  4370. receiving
  4371. [0130 17:51:15 @multigpu.py:323] [p0115]  step: count(869), step_time 1415.04, mean_step_time 1447.47, it/s 0.69
  4372. [0130 17:51:15 @multigpu.py:323] [p0576]  step: count(893), step_time 1428.94, mean_step_time 1444.08, it/s 0.69
  4373. [0130 17:51:15 @multigpu.py:323] [p0574]  step: count(884), step_time 1437.96, mean_step_time 1443.93, it/s 0.69
  4374. sending to address tcp://p0112:61216
  4375. ##### Sending to neptune:  online_score :  0.354189967182 , 1.6 #####
  4376. [u'online', 1.6]
  4377. receiving
  4378. [0130 17:51:16 @multigpu.py:323] [p0576]  step: count(894), step_time 1421.53, mean_step_time 1445.84, it/s 0.69
  4379. [0130 17:51:16 @multigpu.py:323] [p0115]  step: count(870), step_time 1473.59, mean_step_time 1449.05, it/s 0.69
  4380. [0130 17:51:16 @multigpu.py:323] [p0574]  step: count(885), step_time 1468.89, mean_step_time 1444.3, it/s 0.69
  4381. [0130 17:51:18 @multigpu.py:323] [p0576]  step: count(895), step_time 1458.08, mean_step_time 1448.68, it/s 0.69
  4382. [0130 17:51:18 @multigpu.py:323] [p0574]  step: count(886), step_time 1462.1, mean_step_time 1446.39, it/s 0.69
  4383. [0130 17:51:18 @multigpu.py:323] [p0115]  step: count(871), step_time 1488.1, mean_step_time 1453.07, it/s 0.69
  4384. [0130 17:51:19 @multigpu.py:323] [p0576]  step: count(896), step_time 1421.15, mean_step_time 1446.33, it/s 0.69
  4385. [0130 17:51:19 @multigpu.py:323] [p0574]  step: count(887), step_time 1424.94, mean_step_time 1449.1, it/s 0.69
  4386. [0130 17:51:19 @multigpu.py:323] [p0115]  step: count(872), step_time 1470.23, mean_step_time 1456.26, it/s 0.69
  4387. sending to address tcp://p0112:61216
  4388. ##### Sending to neptune:  online_score :  0.35528920273 , 1.4 #####
  4389. [u'online', 1.4]
  4390. receiving
  4391. [0130 17:51:21 @multigpu.py:323] [p0576]  step: count(897), step_time 1400.25, mean_step_time 1443.44, it/s 0.69
  4392. [0130 17:51:21 @multigpu.py:323] [p0574]  step: count(888), step_time 1429.18, mean_step_time 1448.37, it/s 0.69
  4393. [0130 17:51:21 @multigpu.py:323] [p0115]  step: count(873), step_time 1425.93, mean_step_time 1457.84, it/s 0.69
  4394. [0130 17:51:22 @multigpu.py:323] [p0576]  step: count(898), step_time 1511.99, mean_step_time 1447.55, it/s 0.69
  4395. [0130 17:51:22 @multigpu.py:323] [p0574]  step: count(889), step_time 1438.55, mean_step_time 1448.26, it/s 0.69
  4396. [0130 17:51:22 @multigpu.py:323] [p0115]  step: count(874), step_time 1399.34, mean_step_time 1455.56, it/s 0.69
  4397. [0130 17:51:24 @multigpu.py:323] [p0115]  step: count(875), step_time 1390.43, mean_step_time 1451.99, it/s 0.69
  4398. [0130 17:51:24 @multigpu.py:323] [p0574]  step: count(890), step_time 1411.99, mean_step_time 1448.09, it/s 0.69
  4399. [0130 17:51:24 @multigpu.py:323] [p0576]  step: count(899), step_time 1448.49, mean_step_time 1448.48, it/s 0.69
  4400. [0130 17:51:25 @multigpu.py:323] [p0576]  step: count(900), step_time 1443.06, mean_step_time 1446.13, it/s 0.69
  4401. sending debugging info...
  4402. sending to address tcp://p0112:61216
  4403. [0130 17:51:25 @multigpu.py:323] [p0115]  step: count(876), step_time 1472.49, mean_step_time 1453.6, it/s 0.69
  4404. ##### Sending to neptune:  mean_delay :  0.356788568298 , 0.0 #####
  4405. ##### Sending to neptune:  max_delay :  0.356788568298 , -0.0 #####
  4406. sending to address tcp://p0112:61216
  4407. ##### Sending to neptune:  min_delay :  0.356788568298 , -0.0 #####
  4408. [u'delays', [0.0, -0.0, -0.0]]
  4409. receiving
  4410. ##### Sending to neptune:  cost :  0.35678909911 , -0.0122179845348 #####
  4411. ##### Sending to neptune:  policy_loss :  0.35678909911 , -0.376462846994 #####
  4412. sending to address tcp://p0112:61216
  4413. ##### Sending to neptune:  xentropy_loss :  0.35678909911 , -2.28860378265 #####
  4414. ##### Sending to neptune:  value_loss :  0.35678909911 , 1.10116481781 #####
  4415. ##### Sending to neptune:  advantage :  0.35678909911 , 0.00160939083435 #####
  4416. ##### Sending to neptune:  pred_reward :  0.35678909911 , 0.382754832506 #####
  4417. ##### Sending to neptune:  max_logit :  0.35678909911 , 0.197250828147 #####
  4418. [u'loss', -0.012217984534800053, -0.3764628469944, -2.2886037826538086, 1.1011648178100586, 0.0016093908343464136, 0.3827548325061798, 0.1972508281469345]
  4419. receiving
  4420. ##### Sending to neptune:  active_relus :  0.356789570782 , 9117830.05 #####
  4421. ##### Sending to neptune:  dp_per_s :  0.356789570782 , 89.1289708126 #####
  4422. [u'other', 9117830.05, 89.128970812551]
  4423. receiving
  4424. [0130 17:51:25 @multigpu.py:323] [p0574]  step: count(891), step_time 1477.12, mean_step_time 1451.87, it/s 0.69
  4425. [0130 17:51:26 @multigpu.py:323] [p0576]  step: count(901), step_time 1402.28, mean_step_time 1445.38, it/s 0.69
  4426. [0130 17:51:26 @multigpu.py:323] [p0574]  step: count(892), step_time 1390.05, mean_step_time 1449.48, it/s 0.69
  4427. [0130 17:51:26 @multigpu.py:323] [p0115]  step: count(877), step_time 1425.55, mean_step_time 1452.96, it/s 0.69
  4428. [0130 17:51:28 @multigpu.py:323] [p0576]  step: count(902), step_time 1435.78, mean_step_time 1447.48, it/s 0.69
  4429. [0130 17:51:28 @multigpu.py:323] [p0115]  step: count(878), step_time 1411.88, mean_step_time 1451.89, it/s 0.69
  4430. [0130 17:51:28 @multigpu.py:323] [p0574]  step: count(893), step_time 1516.56, mean_step_time 1455.63, it/s 0.69
  4431. [0130 17:51:29 @multigpu.py:323] [p0576]  step: count(903), step_time 1406.34, mean_step_time 1448.72, it/s 0.69
  4432. [0130 17:51:29 @multigpu.py:323] [p0115]  step: count(879), step_time 1462.48, mean_step_time 1453.8, it/s 0.69
  4433. [0130 17:51:29 @multigpu.py:323] [p0574]  step: count(894), step_time 1401.91, mean_step_time 1450.58, it/s 0.69
  4434. [0130 17:51:31 @multigpu.py:323] [p0576]  step: count(904), step_time 1378.58, mean_step_time 1446.5, it/s 0.69
  4435. [0130 17:51:31 @multigpu.py:323] [p0115]  step: count(880), step_time 1418.22, mean_step_time 1454.81, it/s 0.69
  4436. [0130 17:51:31 @multigpu.py:323] [p0574]  step: count(895), step_time 1464.02, mean_step_time 1452.35, it/s 0.69
  4437. [0130 17:51:32 @multigpu.py:323] [p0576]  step: count(905), step_time 1438.57, mean_step_time 1446.68, it/s 0.69
  4438. [0130 17:51:32 @multigpu.py:323] [p0115]  step: count(881), step_time 1422.05, mean_step_time 1454.88, it/s 0.69
  4439. [0130 17:51:32 @multigpu.py:323] [p0574]  step: count(896), step_time 1423.22, mean_step_time 1450.42, it/s 0.69
  4440. [0130 17:51:34 @multigpu.py:323] [p0576]  step: count(906), step_time 1412.81, mean_step_time 1447.98, it/s 0.69
  4441. [0130 17:51:34 @multigpu.py:323] [p0115]  step: count(882), step_time 1423.91, mean_step_time 1453.29, it/s 0.69
  4442. [0130 17:51:34 @multigpu.py:323] [p0574]  step: count(897), step_time 1433.91, mean_step_time 1451.84, it/s 0.69
  4443. [0130 17:51:35 @multigpu.py:323] [p0576]  step: count(907), step_time 1400.85, mean_step_time 1446.11, it/s 0.69
  4444. [0130 17:51:35 @multigpu.py:323] [p0115]  step: count(883), step_time 1509.1, mean_step_time 1458.03, it/s 0.69
  4445. [0130 17:51:35 @multigpu.py:323] [p0574]  step: count(898), step_time 1437.74, mean_step_time 1454.22, it/s 0.69
  4446. [0130 17:51:36 @multigpu.py:323] [p0576]  step: count(908), step_time 1459.62, mean_step_time 1447.92, it/s 0.69
  4447. [0130 17:51:37 @multigpu.py:323] [p0574]  step: count(899), step_time 1388.74, mean_step_time 1449.35, it/s 0.69
  4448. [0130 17:51:37 @multigpu.py:323] [p0115]  step: count(884), step_time 1415.69, mean_step_time 1457.47, it/s 0.69
  4449. sending to address tcp://p0112:61216
  4450. ##### Sending to neptune:  online_score :  0.360088469982 , 0.7 #####
  4451. [u'online', 0.7]
  4452. receiving
  4453. [0130 17:51:38 @multigpu.py:323] [p0576]  step: count(909), step_time 1351.63, mean_step_time 1441.02, it/s 0.69
  4454. [0130 17:51:38 @multigpu.py:323] [p0574]  step: count(900), step_time 1442.32, mean_step_time 1445.03, it/s 0.69
  4455. sending debugging info...
  4456. sending to address tcp://p0112:61216
  4457. ##### Sending to neptune:  mean_delay :  0.360377523568 , 0.0 #####
  4458. sending to address tcp://p0112:61216
  4459. ##### Sending to neptune:  max_delay :  0.360377523568 , -0.0 #####
  4460. ##### Sending to neptune:  min_delay :  0.360377523568 , -0.0 #####
  4461. [u'delays', [0.0, -0.0, -0.0]]
  4462. receiving
  4463. ##### Sending to neptune:  cost :  0.360378028022 , -0.00242157001048 #####
  4464. sending to address tcp://p0112:61216
  4465. ##### Sending to neptune:  policy_loss :  0.360378028022 , 0.664090633392 #####
  4466. ##### Sending to neptune:  xentropy_loss :  0.360378028022 , -2.28888463974 #####
  4467. ##### Sending to neptune:  value_loss :  0.360378028022 , 1.3148329258 #####
  4468. ##### Sending to neptune:  advantage :  0.360378028022 , -0.00289527932182 #####
  4469. ##### Sending to neptune:  pred_reward :  0.360378028022 , 0.36841109395 #####
  4470. ##### Sending to neptune:  max_logit :  0.360378028022 , 0.195855751634 #####
  4471. [u'loss', -0.002421570010483265, 0.664090633392334, -2.2888846397399902, 1.3148329257965088, -0.002895279321819544, 0.3684110939502716, 0.1958557516336441]
  4472. receiving
  4473. ##### Sending to neptune:  active_relus :  0.360378457175 , 9146761.02 #####
  4474. ##### Sending to neptune:  dp_per_s :  0.360378457175 , 88.9845936388 #####
  4475. [u'other', 9146761.02, 88.98459363878979]
  4476. receiving
  4477. [0130 17:51:38 @multigpu.py:323] [p0115]  step: count(885), step_time 1454.51, mean_step_time 1457.07, it/s 0.69
  4478. [0130 17:51:39 @multigpu.py:323] [p0576]  step: count(910), step_time 1534.17, mean_step_time 1446.67, it/s 0.69
  4479. [0130 17:51:39 @multigpu.py:323] [p0574]  step: count(901), step_time 1410.25, mean_step_time 1436.09, it/s 0.7
  4480. [0130 17:51:39 @multigpu.py:323] [p0115]  step: count(886), step_time 1448.48, mean_step_time 1457.54, it/s 0.69
  4481. [0130 17:51:41 @multigpu.py:323] [p0576]  step: count(911), step_time 1396.51, mean_step_time 1442.5, it/s 0.69
  4482. [0130 17:51:41 @multigpu.py:323] [p0574]  step: count(902), step_time 1444.8, mean_step_time 1438.4, it/s 0.7
  4483. [0130 17:51:41 @multigpu.py:323] [p0115]  step: count(887), step_time 1455.82, mean_step_time 1459.59, it/s 0.69
  4484. [0130 17:51:42 @multigpu.py:323] [p0115]  step: count(888), step_time 1403.94, mean_step_time 1439.34, it/s 0.69
  4485. [0130 17:51:42 @multigpu.py:323] [p0574]  step: count(903), step_time 1482.73, mean_step_time 1439.35, it/s 0.69
  4486. [0130 17:51:42 @multigpu.py:323] [p0576]  step: count(912), step_time 1636.16, mean_step_time 1439.34, it/s 0.69
  4487. sending to address tcp://p0112:61216
  4488. ##### Sending to neptune:  online_score :  0.361710860266 , 2.0 #####
  4489. [u'online', 2.0]
  4490. receiving
  4491. [0130 17:51:44 @multigpu.py:323] [p0574]  step: count(904), step_time 1389.31, mean_step_time 1436.92, it/s 0.7
  4492. [0130 17:51:44 @multigpu.py:323] [p0576]  step: count(913), step_time 1450.55, mean_step_time 1440.42, it/s 0.69
  4493. [0130 17:51:44 @multigpu.py:323] [p0115]  step: count(889), step_time 1481.5, mean_step_time 1442.66, it/s 0.69
  4494. [0130 17:51:45 @multigpu.py:323] [p0574]  step: count(905), step_time 1451.54, mean_step_time 1436.05, it/s 0.7
  4495. [0130 17:51:45 @multigpu.py:323] [p0576]  step: count(914), step_time 1420.29, mean_step_time 1440.36, it/s 0.69
  4496. [0130 17:51:45 @multigpu.py:323] [p0115]  step: count(890), step_time 1405.77, mean_step_time 1439.27, it/s 0.69
  4497. [0130 17:51:47 @multigpu.py:323] [p0574]  step: count(906), step_time 1417.06, mean_step_time 1433.8, it/s 0.7
  4498. [0130 17:51:47 @multigpu.py:323] [p0115]  step: count(891), step_time 1381.88, mean_step_time 1433.96, it/s 0.7
  4499. [0130 17:51:47 @multigpu.py:323] [p0576]  step: count(915), step_time 1453.4, mean_step_time 1440.12, it/s 0.69
  4500. [0130 17:51:48 @multigpu.py:323] [p0115]  step: count(892), step_time 1433.03, mean_step_time 1432.1, it/s 0.7
  4501. [0130 17:51:48 @multigpu.py:323] [p0574]  step: count(907), step_time 1453.54, mean_step_time 1435.23, it/s 0.7
  4502. [0130 17:51:48 @multigpu.py:323] [p0576]  step: count(916), step_time 1415.91, mean_step_time 1439.86, it/s 0.69
  4503. [0130 17:51:49 @multigpu.py:323] [p0574]  step: count(908), step_time 1397.95, mean_step_time 1433.67, it/s 0.7
  4504. [0130 17:51:49 @multigpu.py:323] [p0115]  step: count(893), step_time 1417.88, mean_step_time 1431.7, it/s 0.7
  4505. [0130 17:51:49 @multigpu.py:323] [p0576]  step: count(917), step_time 1407.75, mean_step_time 1440.24, it/s 0.69
  4506. sending to address tcp://p0112:61216
  4507. ##### Sending to neptune:  online_score :  0.363701251083 , 1.3 #####
  4508. [u'online', 1.3]
  4509. receiving
  4510. [0130 17:51:51 @multigpu.py:323] [p0574]  step: count(909), step_time 1447.39, mean_step_time 1434.11, it/s 0.7
  4511. [0130 17:51:51 @multigpu.py:323] [p0115]  step: count(894), step_time 1468.22, mean_step_time 1435.14, it/s 0.7
  4512. [0130 17:51:51 @multigpu.py:323] [p0576]  step: count(918), step_time 1449.44, mean_step_time 1437.11, it/s 0.7
  4513. [0130 17:51:52 @multigpu.py:323] [p0574]  step: count(910), step_time 1378.61, mean_step_time 1432.44, it/s 0.7
  4514. [0130 17:51:52 @multigpu.py:323] [p0576]  step: count(919), step_time 1437.36, mean_step_time 1436.55, it/s 0.7
  4515. [0130 17:51:52 @multigpu.py:323] [p0115]  step: count(895), step_time 1470.5, mean_step_time 1439.14, it/s 0.69
  4516. [0130 17:51:54 @multigpu.py:323] [p0574]  step: count(911), step_time 1399.76, mean_step_time 1428.57, it/s 0.7
  4517. [0130 17:51:54 @multigpu.py:323] [p0576]  step: count(920), step_time 1376.39, mean_step_time 1433.22, it/s 0.7
  4518. [0130 17:51:54 @multigpu.py:323] [p0115]  step: count(896), step_time 1374.8, mean_step_time 1434.26, it/s 0.7
  4519. [0130 17:51:55 @multigpu.py:323] [p0574]  step: count(912), step_time 1422.39, mean_step_time 1430.19, it/s 0.7
  4520. [0130 17:51:55 @multigpu.py:323] [p0115]  step: count(897), step_time 1424.27, mean_step_time 1434.2, it/s 0.7
  4521. [0130 17:51:55 @multigpu.py:323] [p0576]  step: count(921), step_time 1496.51, mean_step_time 1437.93, it/s 0.7
  4522. [0130 17:51:56 @multigpu.py:323] [p0574]  step: count(913), step_time 1422.53, mean_step_time 1425.49, it/s 0.7
  4523. [0130 17:51:57 @multigpu.py:323] [p0115]  step: count(898), step_time 1408.28, mean_step_time 1434.02, it/s 0.7
  4524. [0130 17:51:57 @multigpu.py:323] [p0576]  step: count(922), step_time 1424.22, mean_step_time 1437.35, it/s 0.7
  4525. [0130 17:51:58 @multigpu.py:323] [p0574]  step: count(914), step_time 1438.2, mean_step_time 1427.3, it/s 0.7
  4526. [0130 17:51:58 @multigpu.py:323] [p0115]  step: count(899), step_time 1420.96, mean_step_time 1431.94, it/s 0.7
  4527. [0130 17:51:58 @multigpu.py:323] [p0576]  step: count(923), step_time 1398.27, mean_step_time 1436.95, it/s 0.7
  4528. [0130 17:51:59 @multigpu.py:323] [p0574]  step: count(915), step_time 1396.61, mean_step_time 1423.93, it/s 0.7
  4529. [0130 17:51:59 @multigpu.py:323] [p0115]  step: count(900), step_time 1392.29, mean_step_time 1430.64, it/s 0.7
  4530. sending debugging info...
  4531. sending to address tcp://p0112:61216
  4532. ##### Sending to neptune:  mean_delay :  0.366329143312 , 0.0 #####
  4533. sending to address tcp://p0112:61216
  4534. ##### Sending to neptune:  max_delay :  0.366329143312 , -0.0 #####
  4535. ##### Sending to neptune:  min_delay :  0.366329143312 , -0.0 #####
  4536. [u'delays', [0.0, -0.0, -0.0]]
  4537. ##### Sending to neptune:  active_workers :  0.366329351332 , 3 #####
  4538. receiving
  4539. ##### Sending to neptune:  cost :  0.36632966472 , -0.00497338781133 #####
  4540. sending to address tcp://p0112:61216
  4541. ##### Sending to neptune:  policy_loss :  0.36632966472 , 0.409150511026 #####
  4542. ##### Sending to neptune:  xentropy_loss :  0.36632966472 , -2.28915762901 #####
  4543. ##### Sending to neptune:  value_loss :  0.36632966472 , 1.24341356754 #####
  4544. ##### Sending to neptune:  advantage :  0.36632966472 , -0.00184266630094 #####
  4545. ##### Sending to neptune:  pred_reward :  0.36632966472 , 0.362062007189 #####
  4546. ##### Sending to neptune:  max_logit :  0.36632966472 , 0.194479733706 #####
  4547. [u'loss', -0.004973387811332941, 0.40915051102638245, -2.2891576290130615, 1.2434135675430298, -0.001842666300944984, 0.362062007188797, 0.19447973370552063]
  4548. receiving
  4549. ##### Sending to neptune:  active_relus :  0.366330153611 , 9098058.43 #####
  4550. ##### Sending to neptune:  dp_per_s :  0.366330153611 , 88.9767374844 #####
  4551. [u'other', 9098058.43, 88.97673748436563]
  4552. receiving
  4553. [0130 17:51:59 @multigpu.py:323] [p0576]  step: count(924), step_time 1418.48, mean_step_time 1438.94, it/s 0.69
  4554. [0130 17:52:01 @multigpu.py:323] [p0574]  step: count(916), step_time 1421.35, mean_step_time 1423.84, it/s 0.7
  4555. [0130 17:52:01 @multigpu.py:323] [p0115]  step: count(901), step_time 1459.75, mean_step_time 1432.53, it/s 0.7
  4556. [0130 17:52:01 @multigpu.py:323] [p0576]  step: count(925), step_time 1415.02, mean_step_time 1437.77, it/s 0.7
  4557. sending to address tcp://p0112:61216
  4558. ##### Sending to neptune:  online_score :  0.366962843868 , 1.0 #####
  4559. [u'online', 1.0]
  4560. receiving
  4561. [0130 17:52:02 @multigpu.py:323] [p0574]  step: count(917), step_time 1491.58, mean_step_time 1426.72, it/s 0.7
  4562. [0130 17:52:02 @multigpu.py:323] [p0115]  step: count(902), step_time 1418.01, mean_step_time 1432.23, it/s 0.7
  4563. [0130 17:52:02 @multigpu.py:323] [p0576]  step: count(926), step_time 1406.49, mean_step_time 1437.45, it/s 0.7
  4564. [0130 17:52:04 @multigpu.py:323] [p0574]  step: count(918), step_time 1416.31, mean_step_time 1425.65, it/s 0.7
  4565. [0130 17:52:04 @multigpu.py:323] [p0115]  step: count(903), step_time 1426.32, mean_step_time 1428.1, it/s 0.7
  4566. [0130 17:52:04 @multigpu.py:323] [p0576]  step: count(927), step_time 1441.29, mean_step_time 1439.47, it/s 0.69
  4567. [0130 17:52:05 @multigpu.py:323] [p0574]  step: count(919), step_time 1489.08, mean_step_time 1430.67, it/s 0.7
  4568. [0130 17:52:05 @multigpu.py:323] [p0115]  step: count(904), step_time 1470.43, mean_step_time 1430.83, it/s 0.7
  4569. [0130 17:52:05 @multigpu.py:323] [p0576]  step: count(928), step_time 1460.7, mean_step_time 1439.53, it/s 0.69
  4570. [0130 17:52:07 @multigpu.py:323] [p0574]  step: count(920), step_time 1434.11, mean_step_time 1430.25, it/s 0.7
  4571. [0130 17:52:07 @multigpu.py:323] [p0115]  step: count(905), step_time 1428.51, mean_step_time 1429.53, it/s 0.7
  4572. [0130 17:52:07 @multigpu.py:323] [p0576]  step: count(929), step_time 1417.13, mean_step_time 1442.8, it/s 0.69
  4573. [0130 17:52:08 @multigpu.py:323] [p0115]  step: count(906), step_time 1419.18, mean_step_time 1428.07, it/s 0.7
  4574. [0130 17:52:08 @multigpu.py:323] [p0574]  step: count(921), step_time 1438.38, mean_step_time 1431.66, it/s 0.7
  4575. [0130 17:52:08 @multigpu.py:323] [p0576]  step: count(930), step_time 1424.73, mean_step_time 1437.33, it/s 0.7
  4576. [0130 17:52:09 @multigpu.py:323] [p0115]  step: count(907), step_time 1414.07, mean_step_time 1425.98, it/s 0.7
  4577. [0130 17:52:09 @multigpu.py:323] [p0574]  step: count(922), step_time 1411.44, mean_step_time 1429.99, it/s 0.7
  4578. [0130 17:52:09 @multigpu.py:323] [p0576]  step: count(931), step_time 1422.66, mean_step_time 1438.64, it/s 0.7
  4579. sending to address tcp://p0112:61216
  4580. ##### Sending to neptune:  online_score :  0.369363233844 , 1.4 #####
  4581. [u'online', 1.4]
  4582. receiving
  4583. [0130 17:52:11 @multigpu.py:323] [p0115]  step: count(908), step_time 1445.59, mean_step_time 1428.06, it/s 0.7
  4584. [0130 17:52:11 @multigpu.py:323] [p0574]  step: count(923), step_time 1445.31, mean_step_time 1428.12, it/s 0.7
  4585. [0130 17:52:11 @multigpu.py:323] [p0576]  step: count(932), step_time 1426.28, mean_step_time 1428.14, it/s 0.7
  4586. [0130 17:52:12 @multigpu.py:323] [p0115]  step: count(909), step_time 1423.35, mean_step_time 1425.16, it/s 0.7
  4587. [0130 17:52:12 @multigpu.py:323] [p0576]  step: count(933), step_time 1424.03, mean_step_time 1426.82, it/s 0.7
  4588. [0130 17:52:12 @multigpu.py:323] [p0574]  step: count(924), step_time 1471.94, mean_step_time 1432.25, it/s 0.7
  4589. [0130 17:52:14 @multigpu.py:323] [p0115]  step: count(910), step_time 1438.34, mean_step_time 1426.78, it/s 0.7
  4590. [0130 17:52:14 @multigpu.py:323] [p0576]  step: count(934), step_time 1455.56, mean_step_time 1428.58, it/s 0.7
  4591. [0130 17:52:14 @multigpu.py:323] [p0574]  step: count(925), step_time 1450.05, mean_step_time 1432.18, it/s 0.7
  4592. [0130 17:52:15 @multigpu.py:323] [p0115]  step: count(911), step_time 1494.43, mean_step_time 1432.41, it/s 0.7
  4593. [0130 17:52:15 @multigpu.py:323] [p0574]  step: count(926), step_time 1443.2, mean_step_time 1433.49, it/s 0.7
  4594. [0130 17:52:15 @multigpu.py:323] [p0576]  step: count(935), step_time 1490.3, mean_step_time 1430.43, it/s 0.7
  4595. [0130 17:52:17 @multigpu.py:323] [p0576]  step: count(936), step_time 1407.64, mean_step_time 1430.01, it/s 0.7
  4596. [0130 17:52:17 @multigpu.py:323] [p0115]  step: count(912), step_time 1449.7, mean_step_time 1433.24, it/s 0.7
  4597. [0130 17:52:17 @multigpu.py:323] [p0574]  step: count(927), step_time 1445.49, mean_step_time 1433.08, it/s 0.7
  4598. sending to address tcp://p0112:61216
  4599. ##### Sending to neptune:  online_score :  0.371138255795 , 1.8 #####
  4600. [u'online', 1.8]
  4601. receiving
  4602. [0130 17:52:18 @multigpu.py:323] [p0115]  step: count(913), step_time 1401.75, mean_step_time 1432.44, it/s 0.7
  4603. [0130 17:52:18 @multigpu.py:323] [p0576]  step: count(937), step_time 1431.2, mean_step_time 1431.18, it/s 0.7
  4604. [0130 17:52:18 @multigpu.py:323] [p0574]  step: count(928), step_time 1416.73, mean_step_time 1434.02, it/s 0.7
  4605. [0130 17:52:19 @multigpu.py:323] [p0576]  step: count(938), step_time 1408.21, mean_step_time 1429.12, it/s 0.7
  4606. [0130 17:52:20 @multigpu.py:323] [p0574]  step: count(929), step_time 1421.65, mean_step_time 1432.73, it/s 0.7
  4607. [0130 17:52:20 @multigpu.py:323] [p0115]  step: count(914), step_time 1503.67, mean_step_time 1434.21, it/s 0.7
  4608. [0130 17:52:21 @multigpu.py:323] [p0576]  step: count(939), step_time 1432.68, mean_step_time 1428.89, it/s 0.7
  4609. [0130 17:52:21 @multigpu.py:323] [p0115]  step: count(915), step_time 1405.72, mean_step_time 1430.97, it/s 0.7
  4610. [0130 17:52:21 @multigpu.py:323] [p0574]  step: count(930), step_time 1506.38, mean_step_time 1439.12, it/s 0.69
  4611. [0130 17:52:22 @multigpu.py:323] [p0576]  step: count(940), step_time 1390.55, mean_step_time 1429.6, it/s 0.7
  4612. [0130 17:52:22 @multigpu.py:323] [p0574]  step: count(931), step_time 1398.75, mean_step_time 1439.07, it/s 0.69
  4613. [0130 17:52:22 @multigpu.py:323] [p0115]  step: count(916), step_time 1443.18, mean_step_time 1434.39, it/s 0.7
  4614. [0130 17:52:24 @multigpu.py:323] [p0576]  step: count(941), step_time 1413.95, mean_step_time 1425.47, it/s 0.7
  4615. [0130 17:52:24 @multigpu.py:323] [p0115]  step: count(917), step_time 1429.44, mean_step_time 1434.65, it/s 0.7
  4616. [0130 17:52:24 @multigpu.py:323] [p0574]  step: count(932), step_time 1435.86, mean_step_time 1439.75, it/s 0.69
  4617. [0130 17:52:25 @multigpu.py:323] [p0576]  step: count(942), step_time 1419.92, mean_step_time 1425.25, it/s 0.7
  4618. [0130 17:52:25 @multigpu.py:323] [p0574]  step: count(933), step_time 1378.39, mean_step_time 1437.54, it/s 0.7
  4619. [0130 17:52:25 @multigpu.py:323] [p0115]  step: count(918), step_time 1477.73, mean_step_time 1438.12, it/s 0.7
  4620. [0130 17:52:27 @multigpu.py:323] [p0576]  step: count(943), step_time 1432.5, mean_step_time 1426.97, it/s 0.7
  4621. [0130 17:52:27 @multigpu.py:323] [p0574]  step: count(934), step_time 1421.4, mean_step_time 1436.7, it/s 0.7
  4622. [0130 17:52:27 @multigpu.py:323] [p0115]  step: count(919), step_time 1469.72, mean_step_time 1440.56, it/s 0.69
  4623. [0130 17:52:28 @multigpu.py:323] [p0576]  step: count(944), step_time 1463.06, mean_step_time 1429.19, it/s 0.7
  4624. [0130 17:52:28 @multigpu.py:323] [p0574]  step: count(935), step_time 1433.35, mean_step_time 1438.54, it/s 0.7
  4625. [0130 17:52:28 @multigpu.py:323] [p0115]  step: count(920), step_time 1392.17, mean_step_time 1440.55, it/s 0.69
  4626. [0130 17:52:30 @multigpu.py:323] [p0576]  step: count(945), step_time 1464.76, mean_step_time 1431.68, it/s 0.7
  4627. [0130 17:52:30 @multigpu.py:323] [p0574]  step: count(936), step_time 1420.2, mean_step_time 1438.48, it/s 0.7
  4628. [0130 17:52:30 @multigpu.py:323] [p0115]  step: count(921), step_time 1468.92, mean_step_time 1441.01, it/s 0.69
  4629. [0130 17:52:31 @multigpu.py:323] [p0576]  step: count(946), step_time 1380.21, mean_step_time 1430.37, it/s 0.7
  4630. [0130 17:52:31 @multigpu.py:323] [p0574]  step: count(937), step_time 1443.1, mean_step_time 1436.05, it/s 0.7
  4631. [0130 17:52:31 @multigpu.py:323] [p0115]  step: count(922), step_time 1389.22, mean_step_time 1439.57, it/s 0.69
  4632. [0130 17:52:32 @multigpu.py:323] [p0576]  step: count(947), step_time 1413.24, mean_step_time 1428.97, it/s 0.7
  4633. [0130 17:52:32 @multigpu.py:323] [p0574]  step: count(938), step_time 1386.52, mean_step_time 1434.56, it/s 0.7
  4634. [0130 17:52:33 @multigpu.py:323] [p0115]  step: count(923), step_time 1465.42, mean_step_time 1441.53, it/s 0.69
  4635. sending to address tcp://p0112:61216
  4636. ##### Sending to neptune:  online_score :  0.375655125512 , 1.1 #####
  4637. [u'online', 1.1]
  4638. receiving
  4639. [0130 17:52:34 @multigpu.py:323] [p0574]  step: count(939), step_time 1403.78, mean_step_time 1430.3, it/s 0.7
  4640. [0130 17:52:34 @multigpu.py:323] [p0576]  step: count(948), step_time 1467.53, mean_step_time 1429.31, it/s 0.7
  4641. [0130 17:52:34 @multigpu.py:323] [p0115]  step: count(924), step_time 1459.0, mean_step_time 1440.96, it/s 0.69
  4642. [0130 17:52:35 @multigpu.py:323] [p0576]  step: count(949), step_time 1397.18, mean_step_time 1428.31, it/s 0.7
  4643. [0130 17:52:35 @multigpu.py:323] [p0574]  step: count(940), step_time 1420.18, mean_step_time 1429.6, it/s 0.7
  4644. [0130 17:52:35 @multigpu.py:323] [p0115]  step: count(925), step_time 1417.88, mean_step_time 1440.42, it/s 0.69
  4645. [0130 17:52:37 @multigpu.py:323] [p0574]  step: count(941), step_time 1432.8, mean_step_time 1429.32, it/s 0.7
  4646. [0130 17:52:37 @multigpu.py:323] [p0576]  step: count(950), step_time 1466.85, mean_step_time 1430.42, it/s 0.7
  4647. [0130 17:52:37 @multigpu.py:323] [p0115]  step: count(926), step_time 1421.81, mean_step_time 1440.56, it/s 0.69
  4648. [0130 17:52:38 @multigpu.py:323] [p0574]  step: count(942), step_time 1401.24, mean_step_time 1428.81, it/s 0.7
  4649. [0130 17:52:38 @multigpu.py:323] [p0576]  step: count(951), step_time 1443.41, mean_step_time 1431.45, it/s 0.7
  4650. [0130 17:52:38 @multigpu.py:323] [p0115]  step: count(927), step_time 1461.71, mean_step_time 1442.94, it/s 0.69
  4651. [0130 17:52:40 @multigpu.py:323] [p0115]  step: count(928), step_time 1489.4, mean_step_time 1445.13, it/s 0.69
  4652. [0130 17:52:40 @multigpu.py:323] [p0574]  step: count(943), step_time 1767.06, mean_step_time 1444.9, it/s 0.69
  4653. [0130 17:52:40 @multigpu.py:323] [p0576]  step: count(952), step_time 1698.21, mean_step_time 1445.05, it/s 0.69
  4654. sending to address tcp://p0112:61216
  4655. ##### Sending to neptune:  online_score :  0.377704124981 , 1.0 #####
  4656. [u'online', 1.0]
  4657. receiving
  4658. [0130 17:52:41 @multigpu.py:323] [p0115]  step: count(929), step_time 1411.02, mean_step_time 1444.51, it/s 0.69
  4659. [0130 17:52:41 @multigpu.py:323] [p0576]  step: count(953), step_time 1473.86, mean_step_time 1447.54, it/s 0.69
  4660. [0130 17:52:41 @multigpu.py:323] [p0574]  step: count(944), step_time 1499.27, mean_step_time 1446.27, it/s 0.69
  4661. [0130 17:52:43 @multigpu.py:323] [p0115]  step: count(930), step_time 1465.97, mean_step_time 1445.89, it/s 0.69
  4662. [0130 17:52:43 @multigpu.py:323] [p0576]  step: count(954), step_time 1407.2, mean_step_time 1445.12, it/s 0.69
  4663. [0130 17:52:43 @multigpu.py:323] [p0574]  step: count(945), step_time 1412.17, mean_step_time 1444.37, it/s 0.69
  4664. [0130 17:52:44 @multigpu.py:323] [p0576]  step: count(955), step_time 1382.05, mean_step_time 1439.71, it/s 0.69
  4665. [0130 17:52:44 @multigpu.py:323] [p0115]  step: count(931), step_time 1397.76, mean_step_time 1441.06, it/s 0.69
  4666. [0130 17:52:44 @multigpu.py:323] [p0574]  step: count(946), step_time 1457.16, mean_step_time 1445.07, it/s 0.69
  4667. [0130 17:52:46 @multigpu.py:323] [p0115]  step: count(932), step_time 1439.31, mean_step_time 1440.54, it/s 0.69
  4668. [0130 17:52:46 @multigpu.py:323] [p0576]  step: count(956), step_time 1478.99, mean_step_time 1443.28, it/s 0.69
  4669. [0130 17:52:46 @multigpu.py:323] [p0574]  step: count(947), step_time 1416.09, mean_step_time 1443.6, it/s 0.69
  4670. [0130 17:52:47 @multigpu.py:323] [p0115]  step: count(933), step_time 1444.71, mean_step_time 1442.69, it/s 0.69
  4671. [0130 17:52:47 @multigpu.py:323] [p0576]  step: count(957), step_time 1423.76, mean_step_time 1442.91, it/s 0.69
  4672. [0130 17:52:47 @multigpu.py:323] [p0574]  step: count(948), step_time 1437.24, mean_step_time 1444.63, it/s 0.69
  4673. sending to address tcp://p0112:61216
  4674. ##### Sending to neptune:  online_score :  0.379729216364 , 2.1 #####
  4675. [u'online', 2.1]
  4676. receiving
  4677. [0130 17:52:48 @multigpu.py:323] [p0576]  step: count(958), step_time 1396.68, mean_step_time 1442.33, it/s 0.69
  4678. [0130 17:52:48 @multigpu.py:323] [p0115]  step: count(934), step_time 1444.85, mean_step_time 1439.75, it/s 0.69
  4679. [0130 17:52:48 @multigpu.py:323] [p0574]  step: count(949), step_time 1407.62, mean_step_time 1443.93, it/s 0.69
  4680. [0130 17:52:50 @multigpu.py:323] [p0576]  step: count(959), step_time 1417.47, mean_step_time 1441.57, it/s 0.69
  4681. [0130 17:52:50 @multigpu.py:323] [p0574]  step: count(950), step_time 1404.91, mean_step_time 1438.85, it/s 0.69
  4682. [0130 17:52:50 @multigpu.py:323] [p0115]  step: count(935), step_time 1462.43, mean_step_time 1442.58, it/s 0.69
  4683. [0130 17:52:51 @multigpu.py:323] [p0576]  step: count(960), step_time 1401.59, mean_step_time 1442.12, it/s 0.69
  4684. [0130 17:52:51 @multigpu.py:323] [p0574]  step: count(951), step_time 1445.19, mean_step_time 1441.18, it/s 0.69
  4685. [0130 17:52:51 @multigpu.py:323] [p0115]  step: count(936), step_time 1465.44, mean_step_time 1443.7, it/s 0.69
  4686. [0130 17:52:53 @multigpu.py:323] [p0576]  step: count(961), step_time 1413.59, mean_step_time 1442.1, it/s 0.69
  4687. [0130 17:52:53 @multigpu.py:323] [p0574]  step: count(952), step_time 1432.08, mean_step_time 1440.99, it/s 0.69
  4688. [0130 17:52:53 @multigpu.py:323] [p0115]  step: count(937), step_time 1403.75, mean_step_time 1442.41, it/s 0.69
  4689. [0130 17:52:54 @multigpu.py:323] [p0576]  step: count(962), step_time 1436.9, mean_step_time 1442.95, it/s 0.69
  4690. [0130 17:52:54 @multigpu.py:323] [p0574]  step: count(953), step_time 1420.77, mean_step_time 1443.11, it/s 0.69
  4691. [0130 17:52:54 @multigpu.py:323] [p0115]  step: count(938), step_time 1450.15, mean_step_time 1441.03, it/s 0.69
  4692. [0130 17:52:55 @multigpu.py:323] [p0576]  step: count(963), step_time 1391.48, mean_step_time 1440.9, it/s 0.69
  4693. [0130 17:52:56 @multigpu.py:323] [p0115]  step: count(939), step_time 1399.14, mean_step_time 1437.5, it/s 0.7
  4694. [0130 17:52:56 @multigpu.py:323] [p0574]  step: count(954), step_time 1484.84, mean_step_time 1446.28, it/s 0.69
  4695. [0130 17:52:57 @multigpu.py:323] [p0576]  step: count(964), step_time 1456.94, mean_step_time 1440.59, it/s 0.69
  4696. [0130 17:52:57 @multigpu.py:323] [p0115]  step: count(940), step_time 1430.27, mean_step_time 1439.41, it/s 0.69
  4697. [0130 17:52:57 @multigpu.py:323] [p0574]  step: count(955), step_time 1496.64, mean_step_time 1449.44, it/s 0.69
  4698. [0130 17:52:58 @multigpu.py:323] [p0576]  step: count(965), step_time 1422.67, mean_step_time 1438.49, it/s 0.7
  4699. [0130 17:52:58 @multigpu.py:323] [p0115]  step: count(941), step_time 1436.0, mean_step_time 1437.76, it/s 0.7
  4700. [0130 17:52:59 @multigpu.py:323] [p0574]  step: count(956), step_time 1402.66, mean_step_time 1448.57, it/s 0.69
  4701. [0130 17:53:00 @multigpu.py:323] [p0576]  step: count(966), step_time 1394.0, mean_step_time 1439.18, it/s 0.69
  4702. [0130 17:53:00 @multigpu.py:323] [p0115]  step: count(942), step_time 1460.6, mean_step_time 1441.33, it/s 0.69
  4703. [0130 17:53:00 @multigpu.py:323] [p0574]  step: count(957), step_time 1448.67, mean_step_time 1448.84, it/s 0.69
  4704. [0130 17:53:01 @multigpu.py:323] [p0576]  step: count(967), step_time 1405.76, mean_step_time 1438.81, it/s 0.7
  4705. [0130 17:53:01 @multigpu.py:323] [p0115]  step: count(943), step_time 1406.58, mean_step_time 1438.39, it/s 0.7
  4706. [0130 17:53:01 @multigpu.py:323] [p0574]  step: count(958), step_time 1448.4, mean_step_time 1451.94, it/s 0.69
  4707. sending to address tcp://p0112:61216
  4708. ##### Sending to neptune:  online_score :  0.383631636103 , 2.6 #####
  4709. [u'online', 2.6]
  4710. ##### Sending to neptune:  active_workers :  0.383631717165 , 3 #####
  4711. receiving
  4712. [0130 17:53:02 @multigpu.py:323] [p0576]  step: count(968), step_time 1398.77, mean_step_time 1435.37, it/s 0.7
  4713. [0130 17:53:03 @multigpu.py:323] [p0115]  step: count(944), step_time 1453.51, mean_step_time 1438.12, it/s 0.7
  4714. [0130 17:53:03 @multigpu.py:323] [p0574]  step: count(959), step_time 1431.84, mean_step_time 1453.34, it/s 0.69
  4715. [0130 17:53:04 @multigpu.py:323] [p0576]  step: count(969), step_time 1474.77, mean_step_time 1439.25, it/s 0.69
  4716. [0130 17:53:04 @multigpu.py:323] [p0115]  step: count(945), step_time 1394.46, mean_step_time 1436.94, it/s 0.7
  4717. [0130 17:53:04 @multigpu.py:323] [p0574]  step: count(960), step_time 1483.19, mean_step_time 1456.49, it/s 0.69
  4718. [0130 17:53:05 @multigpu.py:323] [p0576]  step: count(970), step_time 1470.58, mean_step_time 1439.43, it/s 0.69
  4719. [0130 17:53:06 @multigpu.py:323] [p0115]  step: count(946), step_time 1519.65, mean_step_time 1441.84, it/s 0.69
  4720. [0130 17:53:06 @multigpu.py:323] [p0574]  step: count(961), step_time 1434.17, mean_step_time 1456.56, it/s 0.69
  4721. [0130 17:53:07 @multigpu.py:323] [p0576]  step: count(971), step_time 1422.48, mean_step_time 1438.39, it/s 0.7
  4722. [0130 17:53:07 @multigpu.py:323] [p0115]  step: count(947), step_time 1403.57, mean_step_time 1438.93, it/s 0.69
  4723. [0130 17:53:07 @multigpu.py:323] [p0574]  step: count(962), step_time 1431.35, mean_step_time 1458.07, it/s 0.69
  4724. sending to address tcp://p0112:61216
  4725. ##### Sending to neptune:  online_score :  0.385261432992 , 2.5 #####
  4726. [u'online', 2.5]
  4727. receiving
  4728. [0130 17:53:09 @multigpu.py:323] [p0574]  step: count(963), step_time 1488.1, mean_step_time 1444.12, it/s 0.69
  4729. [0130 17:53:09 @multigpu.py:323] [p0576]  step: count(972), step_time 1815.85, mean_step_time 1444.27, it/s 0.69
  4730. [0130 17:53:09 @multigpu.py:323] [p0115]  step: count(948), step_time 1599.35, mean_step_time 1444.43, it/s 0.69
  4731. [0130 17:53:10 @multigpu.py:323] [p0115]  step: count(949), step_time 1377.1, mean_step_time 1442.73, it/s 0.69
  4732. [0130 17:53:10 @multigpu.py:323] [p0574]  step: count(964), step_time 1450.41, mean_step_time 1441.67, it/s 0.69
  4733. [0130 17:53:10 @multigpu.py:323] [p0576]  step: count(973), step_time 1457.84, mean_step_time 1443.47, it/s 0.69
  4734. [0130 17:53:11 @multigpu.py:323] [p0115]  step: count(950), step_time 1417.53, mean_step_time 1440.31, it/s 0.69
  4735. [0130 17:53:12 @multigpu.py:323] [p0574]  step: count(965), step_time 1431.62, mean_step_time 1442.65, it/s 0.69
  4736. [0130 17:53:12 @multigpu.py:323] [p0576]  step: count(974), step_time 1460.11, mean_step_time 1446.11, it/s 0.69
  4737. [0130 17:53:13 @multigpu.py:323] [p0115]  step: count(951), step_time 1498.04, mean_step_time 1445.32, it/s 0.69
  4738. [0130 17:53:13 @multigpu.py:323] [p0576]  step: count(975), step_time 1427.4, mean_step_time 1448.38, it/s 0.69
  4739. [0130 17:53:13 @multigpu.py:323] [p0574]  step: count(966), step_time 1468.06, mean_step_time 1443.19, it/s 0.69
  4740. [0130 17:53:14 @multigpu.py:323] [p0115]  step: count(952), step_time 1413.3, mean_step_time 1444.02, it/s 0.69
  4741. [0130 17:53:14 @multigpu.py:323] [p0574]  step: count(967), step_time 1408.96, mean_step_time 1442.84, it/s 0.69
  4742. [0130 17:53:14 @multigpu.py:323] [p0576]  step: count(976), step_time 1427.37, mean_step_time 1445.8, it/s 0.69
  4743. [0130 17:53:16 @multigpu.py:323] [p0115]  step: count(953), step_time 1412.05, mean_step_time 1442.39, it/s 0.69
  4744. [0130 17:53:16 @multigpu.py:323] [p0576]  step: count(977), step_time 1417.17, mean_step_time 1445.47, it/s 0.69
  4745. [0130 17:53:16 @multigpu.py:323] [p0574]  step: count(968), step_time 1450.61, mean_step_time 1443.5, it/s 0.69
  4746. [0130 17:53:17 @multigpu.py:323] [p0574]  step: count(969), step_time 1392.42, mean_step_time 1442.74, it/s 0.69
  4747. [0130 17:53:17 @multigpu.py:323] [p0115]  step: count(954), step_time 1479.82, mean_step_time 1444.14, it/s 0.69
  4748. [0130 17:53:17 @multigpu.py:323] [p0576]  step: count(978), step_time 1437.37, mean_step_time 1447.51, it/s 0.69
  4749. sending to address tcp://p0112:61216
  4750. ##### Sending to neptune:  online_score :  0.388035714693 , 1.1 #####
  4751. [u'online', 1.1]
  4752. receiving
  4753. [0130 17:53:19 @multigpu.py:323] [p0115]  step: count(955), step_time 1388.68, mean_step_time 1440.45, it/s 0.69
  4754. [0130 17:53:19 @multigpu.py:323] [p0574]  step: count(970), step_time 1410.08, mean_step_time 1443.0, it/s 0.69
  4755. [0130 17:53:19 @multigpu.py:323] [p0576]  step: count(979), step_time 1452.04, mean_step_time 1449.23, it/s 0.69
  4756. [0130 17:53:20 @multigpu.py:323] [p0115]  step: count(956), step_time 1440.68, mean_step_time 1439.21, it/s 0.69
  4757. [0130 17:53:20 @multigpu.py:323] [p0574]  step: count(971), step_time 1435.06, mean_step_time 1442.5, it/s 0.69
  4758. [0130 17:53:20 @multigpu.py:323] [p0576]  step: count(980), step_time 1443.09, mean_step_time 1451.31, it/s 0.69
  4759. [0130 17:53:22 @multigpu.py:323] [p0574]  step: count(972), step_time 1401.32, mean_step_time 1440.96, it/s 0.69
  4760. [0130 17:53:22 @multigpu.py:323] [p0115]  step: count(957), step_time 1474.39, mean_step_time 1442.74, it/s 0.69
  4761. [0130 17:53:22 @multigpu.py:323] [p0576]  step: count(981), step_time 1491.56, mean_step_time 1455.21, it/s 0.69
  4762. [0130 17:53:23 @multigpu.py:323] [p0574]  step: count(973), step_time 1395.79, mean_step_time 1439.71, it/s 0.69
  4763. [0130 17:53:23 @multigpu.py:323] [p0115]  step: count(958), step_time 1392.72, mean_step_time 1439.87, it/s 0.69
  4764. [0130 17:53:23 @multigpu.py:323] [p0576]  step: count(982), step_time 1429.28, mean_step_time 1454.83, it/s 0.69
  4765. [0130 17:53:24 @multigpu.py:323] [p0574]  step: count(974), step_time 1448.35, mean_step_time 1437.89, it/s 0.7
  4766. [0130 17:53:24 @multigpu.py:323] [p0115]  step: count(959), step_time 1500.38, mean_step_time 1444.93, it/s 0.69
  4767. [0130 17:53:25 @multigpu.py:323] [p0576]  step: count(983), step_time 1436.56, mean_step_time 1457.08, it/s 0.69
  4768. [0130 17:53:26 @multigpu.py:323] [p0574]  step: count(975), step_time 1406.09, mean_step_time 1433.36, it/s 0.7
  4769. [0130 17:53:26 @multigpu.py:323] [p0115]  step: count(960), step_time 1421.83, mean_step_time 1444.51, it/s 0.69
  4770. [0130 17:53:26 @multigpu.py:323] [p0576]  step: count(984), step_time 1435.02, mean_step_time 1455.99, it/s 0.69
  4771. [0130 17:53:27 @multigpu.py:323] [p0574]  step: count(976), step_time 1450.68, mean_step_time 1435.76, it/s 0.7
  4772. [0130 17:53:27 @multigpu.py:323] [p0115]  step: count(961), step_time 1472.84, mean_step_time 1446.35, it/s 0.69
  4773. [0130 17:53:27 @multigpu.py:323] [p0576]  step: count(985), step_time 1397.26, mean_step_time 1454.72, it/s 0.69
  4774. [0130 17:53:29 @multigpu.py:323] [p0574]  step: count(977), step_time 1470.44, mean_step_time 1436.85, it/s 0.7
  4775. [0130 17:53:29 @multigpu.py:323] [p0576]  step: count(986), step_time 1413.1, mean_step_time 1455.67, it/s 0.69
  4776. [0130 17:53:29 @multigpu.py:323] [p0115]  step: count(962), step_time 1433.59, mean_step_time 1445.0, it/s 0.69
  4777. [0130 17:53:30 @multigpu.py:323] [p0574]  step: count(978), step_time 1425.64, mean_step_time 1435.71, it/s 0.7
  4778. [0130 17:53:30 @multigpu.py:323] [p0576]  step: count(987), step_time 1473.67, mean_step_time 1459.07, it/s 0.69
  4779. [0130 17:53:30 @multigpu.py:323] [p0115]  step: count(963), step_time 1473.48, mean_step_time 1448.35, it/s 0.69
  4780. sending to address tcp://p0112:61216
  4781. ##### Sending to neptune:  online_score :  0.391809822453 , 2.2 #####
  4782. [u'online', 2.2]
  4783. receiving
  4784. [0130 17:53:32 @multigpu.py:323] [p0574]  step: count(979), step_time 1390.61, mean_step_time 1433.65, it/s 0.7
  4785. [0130 17:53:32 @multigpu.py:323] [p0115]  step: count(964), step_time 1384.83, mean_step_time 1444.91, it/s 0.69
  4786. [0130 17:53:32 @multigpu.py:323] [p0576]  step: count(988), step_time 1442.05, mean_step_time 1461.23, it/s 0.68
  4787. [0130 17:53:33 @multigpu.py:323] [p0574]  step: count(980), step_time 1450.56, mean_step_time 1432.02, it/s 0.7
  4788. [0130 17:53:33 @multigpu.py:323] [p0576]  step: count(989), step_time 1396.6, mean_step_time 1457.32, it/s 0.69
  4789. [0130 17:53:33 @multigpu.py:323] [p0115]  step: count(965), step_time 1480.84, mean_step_time 1449.23, it/s 0.69
  4790. [0130 17:53:34 @multigpu.py:323] [p0574]  step: count(981), step_time 1427.71, mean_step_time 1431.69, it/s 0.7
  4791. [0130 17:53:35 @multigpu.py:323] [p0576]  step: count(990), step_time 1409.66, mean_step_time 1454.27, it/s 0.69
  4792. [0130 17:53:35 @multigpu.py:323] [p0115]  step: count(966), step_time 1397.85, mean_step_time 1443.14, it/s 0.69
  4793. [0130 17:53:36 @multigpu.py:323] [p0574]  step: count(982), step_time 1412.82, mean_step_time 1430.77, it/s 0.7
  4794. [0130 17:53:36 @multigpu.py:323] [p0576]  step: count(991), step_time 1453.47, mean_step_time 1455.82, it/s 0.69
  4795. [0130 17:53:36 @multigpu.py:323] [p0115]  step: count(967), step_time 1452.68, mean_step_time 1445.6, it/s 0.69
  4796. [0130 17:53:37 @multigpu.py:323] [p0115]  step: count(968), step_time 1397.95, mean_step_time 1435.53, it/s 0.7
  4797. [0130 17:53:37 @multigpu.py:323] [p0574]  step: count(983), step_time 1587.46, mean_step_time 1435.74, it/s 0.7
  4798. [0130 17:53:37 @multigpu.py:323] [p0576]  step: count(992), step_time 1414.94, mean_step_time 1435.78, it/s 0.7
  4799. [0130 17:53:39 @multigpu.py:323] [p0115]  step: count(969), step_time 1432.59, mean_step_time 1438.3, it/s 0.7
  4800. [0130 17:53:39 @multigpu.py:323] [p0574]  step: count(984), step_time 1438.65, mean_step_time 1435.15, it/s 0.7
  4801. [0130 17:53:39 @multigpu.py:323] [p0576]  step: count(993), step_time 1525.21, mean_step_time 1439.15, it/s 0.69
  4802. sending to address tcp://p0112:61216
  4803. ##### Sending to neptune:  online_score :  0.394126037492 , 1.4 #####
  4804. [u'online', 1.4]
  4805. receiving
  4806. [0130 17:53:40 @multigpu.py:323] [p0115]  step: count(970), step_time 1408.3, mean_step_time 1437.84, it/s 0.7
  4807. [0130 17:53:40 @multigpu.py:323] [p0574]  step: count(985), step_time 1457.66, mean_step_time 1436.45, it/s 0.7
  4808. [0130 17:53:40 @multigpu.py:323] [p0576]  step: count(994), step_time 1441.68, mean_step_time 1438.23, it/s 0.7
  4809. [0130 17:53:42 @multigpu.py:323] [p0115]  step: count(971), step_time 1414.53, mean_step_time 1433.67, it/s 0.7
  4810. [0130 17:53:42 @multigpu.py:323] [p0574]  step: count(986), step_time 1401.51, mean_step_time 1433.12, it/s 0.7
  4811. [0130 17:53:42 @multigpu.py:323] [p0576]  step: count(995), step_time 1420.92, mean_step_time 1437.9, it/s 0.7
  4812. [0130 17:53:43 @multigpu.py:323] [p0115]  step: count(972), step_time 1420.67, mean_step_time 1434.03, it/s 0.7
  4813. [0130 17:53:43 @multigpu.py:323] [p0574]  step: count(987), step_time 1406.8, mean_step_time 1433.01, it/s 0.7
  4814. [0130 17:53:43 @multigpu.py:323] [p0576]  step: count(996), step_time 1412.42, mean_step_time 1437.15, it/s 0.7
  4815. [0130 17:53:45 @multigpu.py:323] [p0115]  step: count(973), step_time 1475.81, mean_step_time 1437.22, it/s 0.7
  4816. [0130 17:53:45 @multigpu.py:323] [p0574]  step: count(988), step_time 1505.52, mean_step_time 1435.76, it/s 0.7
  4817. [0130 17:53:45 @multigpu.py:323] [p0576]  step: count(997), step_time 1430.64, mean_step_time 1437.83, it/s 0.7
  4818. sending to address tcp://p0112:61216
  4819. ##### Sending to neptune:  online_score :  0.395659864677 , 1.1 #####
  4820. [u'online', 1.1]
  4821. receiving
  4822. [0130 17:53:46 @multigpu.py:323] [p0115]  step: count(974), step_time 1447.97, mean_step_time 1435.63, it/s 0.7
  4823. [0130 17:53:46 @multigpu.py:323] [p0574]  step: count(989), step_time 1450.31, mean_step_time 1438.65, it/s 0.7
  4824. [0130 17:53:46 @multigpu.py:323] [p0576]  step: count(998), step_time 1459.58, mean_step_time 1438.94, it/s 0.69
  4825. [0130 17:53:47 @multigpu.py:323] [p0115]  step: count(975), step_time 1414.3, mean_step_time 1436.91, it/s 0.7
  4826. [0130 17:53:47 @multigpu.py:323] [p0574]  step: count(990), step_time 1403.42, mean_step_time 1438.32, it/s 0.7
  4827. [0130 17:53:48 @multigpu.py:323] [p0576]  step: count(999), step_time 1423.1, mean_step_time 1437.49, it/s 0.7
  4828. [0130 17:53:49 @multigpu.py:323] [p0115]  step: count(976), step_time 1433.69, mean_step_time 1436.56, it/s 0.7
  4829. [0130 17:53:49 @multigpu.py:323] [p0576]  step: count(1000), step_time 1402.87, mean_step_time 1435.48, it/s 0.7
  4830. sending debugging info...
  4831. sending to address tcp://p0112:61216
  4832. ##### Sending to neptune:  mean_delay :  0.396757584148 , 0.0 #####
  4833. sending to address tcp://p0112:61216
  4834. ##### Sending to neptune:  max_delay :  0.396757584148 , -0.0 #####
  4835. ##### Sending to neptune:  min_delay :  0.396757584148 , -0.0 #####
  4836. [u'delays', [0.0, -0.0, -0.0]]
  4837. receiving
  4838. ##### Sending to neptune:  cost :  0.396758102775 , -0.00669238669798 #####
  4839. sending to address tcp://p0112:61216
  4840. ##### Sending to neptune:  policy_loss :  0.396758102775 , 0.0405781567097 #####
  4841. ##### Sending to neptune:  xentropy_loss :  0.396758102775 , -2.28896999359 #####
  4842. ##### Sending to neptune:  value_loss :  0.396758102775 , 1.39176607132 #####
  4843. ##### Sending to neptune:  advantage :  0.396758102775 , -0.000199115689611 #####
  4844. ##### Sending to neptune:  pred_reward :  0.396758102775 , 0.394293904305 #####
  4845. ##### Sending to neptune:  max_logit :  0.396758102775 , 0.192680105567 #####
  4846. [u'loss', -0.006692386697977781, 0.04057815670967102, -2.2889699935913086, 1.39176607131958, -0.0001991156896110624, 0.3942939043045044, 0.19268010556697845]
  4847. receiving
  4848. ##### Sending to neptune:  active_relus :  0.396758571664 , 9202785.36 #####
  4849. ##### Sending to neptune:  dp_per_s :  0.396758571664 , 88.8521250407 #####
  4850. [u'other', 9202785.36, 88.8521250406519]
  4851. receiving
  4852. 100%|##########|1000/1000[23:48<00:00, 0.70it/s]
  4853. [0130 17:53:49 @multigpu.py:323] [p0574]  step: count(991), step_time 1493.14, mean_step_time 1441.22, it/s 0.69
  4854. [0130 17:53:49 @stat.py:81] async_global_step: 1000
  4855. [0130 17:53:49 @stat.py:81] learning_rate_1: 0.00015
  4856. EPOCH ENDS HERE
  4857. [0130 17:53:49 @timer.py:52] Epoch 1, global_step=1000 finished, time=1428.18sec.
  4858. [0130 17:53:50 @multigpu.py:323] [p0115]  step: count(977), step_time 1437.93, mean_step_time 1434.74, it/s 0.7
  4859. [0130 17:53:50 @multigpu.py:323] [p0574]  step: count(992), step_time 1422.59, mean_step_time 1442.29, it/s 0.69
  4860. [0130 17:53:50 @multigpu.py:323] [p0576]  step: count(1001), step_time 1473.44, mean_step_time 1434.58, it/s 0.7
  4861. [0130 17:53:52 @multigpu.py:323] [p0115]  step: count(978), step_time 1459.16, mean_step_time 1438.06, it/s 0.7
  4862. [0130 17:53:52 @multigpu.py:323] [p0576]  step: count(1002), step_time 1379.25, mean_step_time 1432.07, it/s 0.7
  4863. [0130 17:53:52 @multigpu.py:323] [p0574]  step: count(993), step_time 1417.62, mean_step_time 1443.38, it/s 0.69
  4864. [0130 17:53:53 @multigpu.py:323] [p0115]  step: count(979), step_time 1421.27, mean_step_time 1434.11, it/s 0.7
  4865. [0130 17:53:53 @multigpu.py:323] [p0576]  step: count(1003), step_time 1409.38, mean_step_time 1430.71, it/s 0.7
  4866. [0130 17:53:53 @multigpu.py:323] [p0574]  step: count(994), step_time 1445.48, mean_step_time 1443.24, it/s 0.69
  4867. [0130 17:53:55 @multigpu.py:323] [p0576]  step: count(1004), step_time 1420.49, mean_step_time 1429.99, it/s 0.7
  4868. [0130 17:53:55 @multigpu.py:323] [p0115]  step: count(980), step_time 1438.41, mean_step_time 1434.93, it/s 0.7
  4869. [0130 17:53:55 @multigpu.py:323] [p0574]  step: count(995), step_time 1407.12, mean_step_time 1443.29, it/s 0.69
  4870. [0130 17:53:56 @multigpu.py:323] [p0576]  step: count(1005), step_time 1421.42, mean_step_time 1431.2, it/s 0.7
  4871. [0130 17:53:56 @multigpu.py:323] [p0115]  step: count(981), step_time 1457.56, mean_step_time 1434.17, it/s 0.7
  4872. [0130 17:53:56 @multigpu.py:323] [p0574]  step: count(996), step_time 1433.5, mean_step_time 1442.43, it/s 0.69
  4873. [0130 17:53:58 @multigpu.py:323] [p0576]  step: count(1006), step_time 1467.35, mean_step_time 1433.91, it/s 0.7
  4874. [0130 17:53:58 @multigpu.py:323] [p0574]  step: count(997), step_time 1414.08, mean_step_time 1439.61, it/s 0.69
  4875. [0130 17:53:58 @multigpu.py:323] [p0115]  step: count(982), step_time 1437.83, mean_step_time 1434.38, it/s 0.7
  4876. [0130 17:53:59 @multigpu.py:323] [p0576]  step: count(1007), step_time 1384.39, mean_step_time 1429.44, it/s 0.7
  4877. [0130 17:53:59 @multigpu.py:323] [p0115]  step: count(983), step_time 1396.47, mean_step_time 1430.53, it/s 0.7
  4878. [0130 17:53:59 @multigpu.py:323] [p0574]  step: count(998), step_time 1408.16, mean_step_time 1438.74, it/s 0.7
  4879. [0130 17:54:00 @multigpu.py:323] [p0576]  step: count(1008), step_time 1379.76, mean_step_time 1426.33, it/s 0.7
  4880. [0130 17:54:00 @multigpu.py:323] [p0115]  step: count(984), step_time 1402.52, mean_step_time 1431.42, it/s 0.7
  4881. [0130 17:54:00 @multigpu.py:323] [p0574]  step: count(999), step_time 1406.89, mean_step_time 1439.55, it/s 0.69
  4882. sending to address tcp://p0112:61216
  4883. ##### Sending to neptune:  online_score :  0.400034431352 , 0.8 #####
  4884. [u'online', 0.8]
  4885. receiving
  4886. [0130 17:54:02 @multigpu.py:323] [p0576]  step: count(1009), step_time 1431.42, mean_step_time 1428.07, it/s 0.7
  4887. [0130 17:54:02 @multigpu.py:323] [p0574]  step: count(1000), step_time 1397.69, mean_step_time 1436.91, it/s 0.7
  4888. sending debugging info...
  4889. sending to address tcp://p0112:61216
  4890. ##### Sending to neptune:  mean_delay :  0.400312784976 , 0.0 #####
  4891. ##### Sending to neptune:  max_delay :  0.400312784976 , -0.0 #####
  4892. sending to address tcp://p0112:61216
  4893. ##### Sending to neptune:  min_delay :  0.400312784976 , -0.0 #####
  4894. [u'delays', [0.0, -0.0, -0.0]]
  4895. ##### Sending to neptune:  active_workers :  0.400312963592 , 3 #####
  4896. receiving
  4897. ##### Sending to neptune:  cost :  0.400313278304 , -0.014850821346 #####
  4898. ##### Sending to neptune:  policy_loss :  0.400313278304 , -0.792145967484 #####
  4899. sending to address tcp://p0112:61216
  4900. ##### Sending to neptune:  xentropy_loss :  0.400313278304 , -2.28885960579 #####
  4901. ##### Sending to neptune:  value_loss :  0.400313278304 , 1.18010020256 #####
  4902. ##### Sending to neptune:  advantage :  0.400313278304 , 0.00343013252132 #####
  4903. ##### Sending to neptune:  pred_reward :  0.400313278304 , 0.402852445841 #####
  4904. ##### Sending to neptune:  max_logit :  0.400313278304 , 0.192205965519 #####
  4905. [u'loss', -0.01485082134604454, -0.7921459674835205, -2.2888596057891846, 1.1801002025604248, 0.003430132521316409, 0.40285244584083557, 0.19220596551895142]
  4906. receiving
  4907. ##### Sending to neptune:  active_relus :  0.400313773884 , 9198748.82 #####
  4908. ##### Sending to neptune:  dp_per_s :  0.400313773884 , 89.0196620523 #####
  4909. [u'other', 9198748.82, 89.0196620522503]
  4910. receiving
  4911. 100%|##########|1000/1000[24:00<00:00, 0.69it/s]
  4912. [0130 17:54:02 @stat.py:81] async_global_step: 1000
  4913. [0130 17:54:02 @stat.py:81] learning_rate_1: 0.00015
  4914. EPOCH ENDS HERE
  4915. [0130 17:54:02 @timer.py:52] Epoch 1, global_step=1000 finished, time=1440.88sec.
  4916. [0130 17:54:02 @multigpu.py:323] [p0115]  step: count(985), step_time 1469.01, mean_step_time 1430.82, it/s 0.7
  4917. [0130 17:54:03 @multigpu.py:323] [p0576]  step: count(1010), step_time 1390.46, mean_step_time 1427.11, it/s 0.7
  4918. [0130 17:54:03 @multigpu.py:323] [p0574]  step: count(1001), step_time 1500.74, mean_step_time 1440.56, it/s 0.69
  4919. [0130 17:54:03 @multigpu.py:323] [p0115]  step: count(986), step_time 1467.34, mean_step_time 1434.3, it/s 0.7
  4920. [0130 17:54:04 @multigpu.py:323] [p0576]  step: count(1011), step_time 1380.93, mean_step_time 1423.48, it/s 0.7
  4921. [0130 17:54:05 @multigpu.py:323] [p0574]  step: count(1002), step_time 1482.45, mean_step_time 1444.04, it/s 0.69
  4922. [0130 17:54:05 @multigpu.py:323] [p0115]  step: count(987), step_time 1484.52, mean_step_time 1435.89, it/s 0.7
  4923. sending to address tcp://p0112:61216
  4924. ##### Sending to neptune:  online_score :  0.40136244052 , 1.6 #####
  4925. [u'online', 1.6]
  4926. receiving
  4927. [0130 17:54:06 @multigpu.py:323] [p0115]  step: count(988), step_time 1423.16, mean_step_time 1437.15, it/s 0.7
  4928. [0130 17:54:06 @multigpu.py:323] [p0574]  step: count(1003), step_time 1449.53, mean_step_time 1437.14, it/s 0.7
  4929. [0130 17:54:06 @multigpu.py:323] [p0576]  step: count(1012), step_time 1688.2, mean_step_time 1437.15, it/s 0.7
  4930. sending to address tcp://p0112:61216
  4931. ##### Sending to neptune:  online_score :  0.40154918472 , 0.9 #####
  4932. [u'online', 0.9]
  4933. receiving
  4934. [0130 17:54:08 @multigpu.py:323] [p0574]  step: count(1004), step_time 1386.11, mean_step_time 1434.52, it/s 0.7
  4935. [0130 17:54:08 @multigpu.py:323] [p0115]  step: count(989), step_time 1424.58, mean_step_time 1436.75, it/s 0.7
  4936. [0130 17:54:08 @multigpu.py:323] [p0576]  step: count(1013), step_time 1437.21, mean_step_time 1432.75, it/s 0.7
  4937. [0130 17:54:09 @multigpu.py:323] [p0576]  step: count(1014), step_time 1386.41, mean_step_time 1429.98, it/s 0.7
  4938. [0130 17:54:09 @multigpu.py:323] [p0574]  step: count(1005), step_time 1472.24, mean_step_time 1435.25, it/s 0.7
  4939. [0130 17:54:09 @multigpu.py:323] [p0115]  step: count(990), step_time 1449.78, mean_step_time 1438.83, it/s 0.7
  4940. [0130 17:54:10 @multigpu.py:323] [p0574]  step: count(1006), step_time 1416.46, mean_step_time 1435.99, it/s 0.7
  4941. [0130 17:54:10 @multigpu.py:323] [p0576]  step: count(1015), step_time 1453.32, mean_step_time 1431.6, it/s 0.7
  4942. [0130 17:54:10 @multigpu.py:323] [p0115]  step: count(991), step_time 1406.55, mean_step_time 1438.43, it/s 0.7
  4943. [0130 17:54:12 @multigpu.py:323] [p0576]  step: count(1016), step_time 1376.92, mean_step_time 1429.83, it/s 0.7
  4944. [0130 17:54:12 @multigpu.py:323] [p0574]  step: count(1007), step_time 1404.55, mean_step_time 1435.88, it/s 0.7
  4945. [0130 17:54:12 @multigpu.py:323] [p0115]  step: count(992), step_time 1434.62, mean_step_time 1439.12, it/s 0.69
  4946. [0130 17:54:13 @multigpu.py:323] [p0574]  step: count(1008), step_time 1392.2, mean_step_time 1430.21, it/s 0.7
  4947. [0130 17:54:13 @multigpu.py:323] [p0115]  step: count(993), step_time 1443.97, mean_step_time 1437.53, it/s 0.7
  4948. [0130 17:54:13 @multigpu.py:323] [p0576]  step: count(1017), step_time 1511.07, mean_step_time 1433.85, it/s 0.7
  4949. [0130 17:54:15 @multigpu.py:323] [p0574]  step: count(1009), step_time 1434.25, mean_step_time 1429.41, it/s 0.7
  4950. [0130 17:54:15 @multigpu.py:323] [p0115]  step: count(994), step_time 1373.23, mean_step_time 1433.8, it/s 0.7
  4951. [0130 17:54:15 @multigpu.py:323] [p0576]  step: count(1018), step_time 1411.65, mean_step_time 1431.45, it/s 0.7
  4952. [0130 17:54:16 @multigpu.py:323] [p0574]  step: count(1010), step_time 1381.34, mean_step_time 1428.31, it/s 0.7
  4953. [0130 17:54:16 @multigpu.py:323] [p0115]  step: count(995), step_time 1433.91, mean_step_time 1434.78, it/s 0.7
  4954. [0130 17:54:16 @multigpu.py:323] [p0576]  step: count(1019), step_time 1424.85, mean_step_time 1431.54, it/s 0.7
  4955. [0130 17:54:17 @multigpu.py:323] [p0574]  step: count(1011), step_time 1408.96, mean_step_time 1424.1, it/s 0.7
  4956. [0130 17:54:18 @multigpu.py:323] [p0115]  step: count(996), step_time 1402.86, mean_step_time 1433.23, it/s 0.7
  4957. [0130 17:54:18 @multigpu.py:323] [p0576]  step: count(1020), step_time 1480.01, mean_step_time 1435.4, it/s 0.7
  4958. [0130 17:54:19 @multigpu.py:323] [p0574]  step: count(1012), step_time 1432.06, mean_step_time 1424.57, it/s 0.7
  4959. [0130 17:54:19 @multigpu.py:323] [p0115]  step: count(997), step_time 1412.77, mean_step_time 1431.98, it/s 0.7
  4960. [0130 17:54:19 @multigpu.py:323] [p0576]  step: count(1021), step_time 1444.97, mean_step_time 1433.97, it/s 0.7
  4961. [0130 17:54:20 @multigpu.py:323] [p0574]  step: count(1013), step_time 1448.9, mean_step_time 1426.13, it/s 0.7
  4962. [0130 17:54:20 @multigpu.py:323] [p0115]  step: count(998), step_time 1395.76, mean_step_time 1428.81, it/s 0.7
  4963. [0130 17:54:21 @multigpu.py:323] [p0576]  step: count(1022), step_time 1410.01, mean_step_time 1435.51, it/s 0.7
  4964. [0130 17:54:22 @multigpu.py:323] [p0574]  step: count(1014), step_time 1413.5, mean_step_time 1424.54, it/s 0.7
  4965. [0130 17:54:22 @multigpu.py:323] [p0115]  step: count(999), step_time 1438.66, mean_step_time 1429.68, it/s 0.7
  4966. [0130 17:54:22 @multigpu.py:323] [p0576]  step: count(1023), step_time 1382.26, mean_step_time 1434.16, it/s 0.7
  4967. [0130 17:54:23 @multigpu.py:323] [p0574]  step: count(1015), step_time 1425.44, mean_step_time 1425.45, it/s 0.7
  4968. [0130 17:54:23 @multigpu.py:323] [p0115]  step: count(1000), step_time 1496.72, mean_step_time 1432.59, it/s 0.7
  4969. sending debugging info...
  4970. sending to address tcp://p0112:61216
  4971. ##### Sending to neptune:  mean_delay :  0.406299769415 , 0.0 #####
  4972. sending to address tcp://p0112:61216
  4973. ##### Sending to neptune:  max_delay :  0.406299769415 , -0.0 #####
  4974. ##### Sending to neptune:  min_delay :  0.406299769415 , -0.0 #####
  4975. [u'delays', [0.0, -0.0, -0.0]]
  4976. receiving
  4977. ##### Sending to neptune:  cost :  0.406300278306 , -0.00499816797674 #####
  4978. sending to address tcp://p0112:61216
  4979. ##### Sending to neptune:  policy_loss :  0.406300278306 , 0.282140910625 #####
  4980. ##### Sending to neptune:  xentropy_loss :  0.406300278306 , -2.2886068821 #####
  4981. ##### Sending to neptune:  value_loss :  0.406300278306 , 1.36670041084 #####
  4982. ##### Sending to neptune:  advantage :  0.406300278306 , -0.00127228489146 #####
  4983. ##### Sending to neptune:  pred_reward :  0.406300278306 , 0.407911151648 #####
  4984. ##### Sending to neptune:  max_logit :  0.406300278306 , 0.192866250873 #####
  4985. [u'loss', -0.004998167976737022, 0.28214091062545776, -2.288606882095337, 1.3667004108428955, -0.0012722848914563656, 0.40791115164756775, 0.192866250872612]
  4986. 100%|##########|1000/1000[23:53<00:00, 0.70it/s]
  4987. receiving
  4988. ##### Sending to neptune:  active_relus :  0.406300744679 , 9248289.22 #####
  4989. ##### Sending to neptune:  dp_per_s :  0.406300744679 , 89.0294158633 #####
  4990. [u'other', 9248289.22, 89.02941586332658]
  4991. receiving
  4992. [0130 17:54:23 @stat.py:81] async_global_step: 1000
  4993. [0130 17:54:23 @stat.py:81] learning_rate_1: 0.00015
  4994. EPOCH ENDS HERE
  4995. [0130 17:54:23 @timer.py:52] Epoch 1, global_step=1000 finished, time=1433.06sec.
  4996. [0130 17:54:23 @multigpu.py:323] [p0576]  step: count(1024), step_time 1482.16, mean_step_time 1437.24, it/s 0.7
  4997. sending to address tcp://p0112:61216
  4998. ##### Sending to neptune:  online_score :  0.406547241608 , 1.9 #####
  4999. [u'online', 1.9]
  5000. receiving
  5001. [0130 17:54:25 @multigpu.py:323] [p0574]  step: count(1016), step_time 1396.65, mean_step_time 1423.61, it/s 0.7
  5002. [0130 17:54:25 @multigpu.py:323] [p0115]  step: count(1001), step_time 1433.23, mean_step_time 1431.37, it/s 0.7
  5003. [0130 17:54:25 @multigpu.py:323] [p0576]  step: count(1025), step_time 1436.77, mean_step_time 1438.01, it/s 0.7
  5004. [0130 17:54:26 @multigpu.py:323] [p0574]  step: count(1017), step_time 1386.37, mean_step_time 1422.22, it/s 0.7
  5005. [0130 17:54:26 @multigpu.py:323] [p0115]  step: count(1002), step_time 1394.99, mean_step_time 1429.23, it/s 0.7
  5006. [0130 17:54:26 @multigpu.py:323] [p0576]  step: count(1026), step_time 1390.23, mean_step_time 1434.15, it/s 0.7
  5007. [0130 17:54:27 @multigpu.py:323] [p0574]  step: count(1018), step_time 1411.35, mean_step_time 1422.38, it/s 0.7
  5008. [0130 17:54:28 @multigpu.py:323] [p0115]  step: count(1003), step_time 1436.11, mean_step_time 1431.21, it/s 0.7
  5009. [0130 17:54:28 @multigpu.py:323] [p0576]  step: count(1027), step_time 1417.47, mean_step_time 1435.8, it/s 0.7
  5010. [0130 17:54:29 @multigpu.py:323] [p0574]  step: count(1019), step_time 1413.31, mean_step_time 1422.7, it/s 0.7
  5011. [0130 17:54:29 @multigpu.py:323] [p0115]  step: count(1004), step_time 1418.69, mean_step_time 1432.02, it/s 0.7
  5012. [0130 17:54:29 @multigpu.py:323] [p0576]  step: count(1028), step_time 1393.45, mean_step_time 1436.49, it/s 0.7
  5013. [0130 17:54:30 @multigpu.py:323] [p0574]  step: count(1020), step_time 1430.03, mean_step_time 1424.32, it/s 0.7
  5014. [0130 17:54:30 @multigpu.py:323] [p0115]  step: count(1005), step_time 1461.79, mean_step_time 1431.66, it/s 0.7
  5015. [0130 17:54:30 @multigpu.py:323] [p0576]  step: count(1029), step_time 1430.71, mean_step_time 1436.45, it/s 0.7
  5016. [0130 17:54:32 @multigpu.py:323] [p0574]  step: count(1021), step_time 1470.66, mean_step_time 1422.82, it/s 0.7
  5017. [0130 17:54:32 @multigpu.py:323] [p0115]  step: count(1006), step_time 1423.65, mean_step_time 1429.48, it/s 0.7
  5018. [0130 17:54:32 @multigpu.py:323] [p0576]  step: count(1030), step_time 1439.9, mean_step_time 1438.93, it/s 0.69
  5019. [0130 17:54:33 @multigpu.py:323] [p0574]  step: count(1022), step_time 1408.78, mean_step_time 1419.13, it/s 0.7
  5020. [0130 17:54:33 @multigpu.py:323] [p0115]  step: count(1007), step_time 1421.95, mean_step_time 1426.35, it/s 0.7
  5021. [0130 17:54:33 @multigpu.py:323] [p0576]  step: count(1031), step_time 1445.41, mean_step_time 1442.15, it/s 0.69
  5022. [0130 17:54:35 @multigpu.py:323] [p0576]  step: count(1032), step_time 1395.33, mean_step_time 1427.51, it/s 0.7
  5023. [0130 17:54:35 @multigpu.py:323] [p0574]  step: count(1023), step_time 1614.89, mean_step_time 1427.4, it/s 0.7
  5024. [0130 17:54:35 @multigpu.py:323] [p0115]  step: count(1008), step_time 1446.89, mean_step_time 1427.54, it/s 0.7
  5025. [0130 17:54:36 @multigpu.py:323] [p0574]  step: count(1024), step_time 1390.41, mean_step_time 1427.62, it/s 0.7
  5026. sending to address tcp://p0112:61216
  5027. ##### Sending to neptune:  online_score :  0.409867774116 , 2.2 #####
  5028. [u'online', 2.2]
  5029. receiving
  5030. [0130 17:54:36 @multigpu.py:323] [p0576]  step: count(1033), step_time 1416.02, mean_step_time 1426.45, it/s 0.7
  5031. [0130 17:54:36 @multigpu.py:323] [p0115]  step: count(1009), step_time 1435.11, mean_step_time 1428.06, it/s 0.7
  5032. [0130 17:54:38 @multigpu.py:323] [p0115]  step: count(1010), step_time 1396.53, mean_step_time 1425.4, it/s 0.7
  5033. [0130 17:54:38 @multigpu.py:323] [p0574]  step: count(1025), step_time 1454.39, mean_step_time 1426.72, it/s 0.7
  5034. [0130 17:54:38 @multigpu.py:323] [p0576]  step: count(1034), step_time 1454.06, mean_step_time 1429.83, it/s 0.7
  5035. sending to address tcp://p0112:61216
  5036. ##### Sending to neptune:  online_score :  0.410494857497 , 1.8 #####
  5037. [u'online', 1.8]
  5038. receiving
  5039. [0130 17:54:39 @multigpu.py:323] [p0115]  step: count(1011), step_time 1457.56, mean_step_time 1427.95, it/s 0.7
  5040. [0130 17:54:39 @multigpu.py:323] [p0574]  step: count(1026), step_time 1464.19, mean_step_time 1429.11, it/s 0.7
  5041. [0130 17:54:39 @multigpu.py:323] [p0576]  step: count(1035), step_time 1483.78, mean_step_time 1431.35, it/s 0.7
  5042. [0130 17:54:40 @multigpu.py:323] [p0115]  step: count(1012), step_time 1456.61, mean_step_time 1429.05, it/s 0.7
  5043. [0130 17:54:40 @multigpu.py:323] [p0574]  step: count(1027), step_time 1441.26, mean_step_time 1430.95, it/s 0.7
  5044. [0130 17:54:41 @multigpu.py:323] [p0576]  step: count(1036), step_time 1423.45, mean_step_time 1433.68, it/s 0.7
  5045. [0130 17:54:42 @multigpu.py:323] [p0574]  step: count(1028), step_time 1432.02, mean_step_time 1432.94, it/s 0.7
  5046. [0130 17:54:42 @multigpu.py:323] [p0115]  step: count(1013), step_time 1439.95, mean_step_time 1428.85, it/s 0.7
  5047. [0130 17:54:42 @multigpu.py:323] [p0576]  step: count(1037), step_time 1412.17, mean_step_time 1428.73, it/s 0.7
  5048. [0130 17:54:43 @multigpu.py:323] [p0115]  step: count(1014), step_time 1422.85, mean_step_time 1431.33, it/s 0.7
  5049. [0130 17:54:43 @multigpu.py:323] [p0576]  step: count(1038), step_time 1440.15, mean_step_time 1430.16, it/s 0.7
  5050. [0130 17:54:43 @multigpu.py:323] [p0574]  step: count(1029), step_time 1458.63, mean_step_time 1434.16, it/s 0.7
  5051. [0130 17:54:45 @multigpu.py:323] [p0576]  step: count(1039), step_time 1383.62, mean_step_time 1428.1, it/s 0.7
  5052. [0130 17:54:45 @multigpu.py:323] [p0115]  step: count(1015), step_time 1440.15, mean_step_time 1431.64, it/s 0.7
  5053. [0130 17:54:45 @multigpu.py:323] [p0574]  step: count(1030), step_time 1436.54, mean_step_time 1436.92, it/s 0.7
  5054. [0130 17:54:46 @multigpu.py:323] [p0576]  step: count(1040), step_time 1419.95, mean_step_time 1425.09, it/s 0.7
  5055. [0130 17:54:46 @multigpu.py:323] [p0574]  step: count(1031), step_time 1453.02, mean_step_time 1439.12, it/s 0.69
  5056. [0130 17:54:46 @multigpu.py:323] [p0115]  step: count(1016), step_time 1500.55, mean_step_time 1436.53, it/s 0.7
  5057. [0130 17:54:48 @multigpu.py:323] [p0576]  step: count(1041), step_time 1415.08, mean_step_time 1423.6, it/s 0.7
  5058. [0130 17:54:48 @multigpu.py:323] [p0574]  step: count(1032), step_time 1400.81, mean_step_time 1437.56, it/s 0.7
  5059. [0130 17:54:48 @multigpu.py:323] [p0115]  step: count(1017), step_time 1412.98, mean_step_time 1436.54, it/s 0.7
  5060. [0130 17:54:49 @multigpu.py:323] [p0576]  step: count(1042), step_time 1366.53, mean_step_time 1421.42, it/s 0.7
  5061. [0130 17:54:49 @multigpu.py:323] [p0574]  step: count(1033), step_time 1453.49, mean_step_time 1437.79, it/s 0.7
  5062. [0130 17:54:49 @multigpu.py:323] [p0115]  step: count(1018), step_time 1435.46, mean_step_time 1438.52, it/s 0.7
  5063. [0130 17:54:50 @multigpu.py:323] [p0576]  step: count(1043), step_time 1474.24, mean_step_time 1426.02, it/s 0.7
  5064. [0130 17:54:51 @multigpu.py:323] [p0574]  step: count(1034), step_time 1398.87, mean_step_time 1437.06, it/s 0.7
  5065. [0130 17:54:51 @multigpu.py:323] [p0115]  step: count(1019), step_time 1426.7, mean_step_time 1437.92, it/s 0.7
  5066. [0130 17:54:52 @multigpu.py:323] [p0576]  step: count(1044), step_time 1480.67, mean_step_time 1425.95, it/s 0.7
  5067. [0130 17:54:52 @multigpu.py:323] [p0574]  step: count(1035), step_time 1418.28, mean_step_time 1436.7, it/s 0.7
  5068. [0130 17:54:52 @multigpu.py:323] [p0115]  step: count(1020), step_time 1414.59, mean_step_time 1433.82, it/s 0.7
  5069. [0130 17:54:53 @multigpu.py:323] [p0574]  step: count(1036), step_time 1373.41, mean_step_time 1435.54, it/s 0.7
  5070. [0130 17:54:53 @multigpu.py:323] [p0576]  step: count(1045), step_time 1500.63, mean_step_time 1429.14, it/s 0.7
  5071. [0130 17:54:53 @multigpu.py:323] [p0115]  step: count(1021), step_time 1447.12, mean_step_time 1434.51, it/s 0.7
  5072. sending to address tcp://p0112:61216
  5073. ##### Sending to neptune:  online_score :  0.414848094715 , 1.4 #####
  5074. [u'online', 1.4]
  5075. receiving
  5076. [0130 17:54:55 @multigpu.py:323] [p0574]  step: count(1037), step_time 1446.01, mean_step_time 1438.52, it/s 0.7
  5077. [0130 17:54:55 @multigpu.py:323] [p0576]  step: count(1046), step_time 1383.97, mean_step_time 1428.83, it/s 0.7
  5078. [0130 17:54:55 @multigpu.py:323] [p0115]  step: count(1022), step_time 1451.57, mean_step_time 1437.34, it/s 0.7
  5079. [0130 17:54:56 @multigpu.py:323] [p0574]  step: count(1038), step_time 1386.57, mean_step_time 1437.28, it/s 0.7
  5080. [0130 17:54:56 @multigpu.py:323] [p0576]  step: count(1047), step_time 1478.74, mean_step_time 1431.89, it/s 0.7
  5081. [0130 17:54:56 @multigpu.py:323] [p0115]  step: count(1023), step_time 1435.33, mean_step_time 1437.3, it/s 0.7
  5082. [0130 17:54:58 @multigpu.py:323] [p0574]  step: count(1039), step_time 1427.69, mean_step_time 1438.0, it/s 0.7
  5083. [0130 17:54:58 @multigpu.py:323] [p0576]  step: count(1048), step_time 1418.64, mean_step_time 1433.15, it/s 0.7
  5084. [0130 17:54:58 @multigpu.py:323] [p0115]  step: count(1024), step_time 1436.89, mean_step_time 1438.21, it/s 0.7
  5085. [0130 17:54:59 @multigpu.py:323] [p0574]  step: count(1040), step_time 1395.55, mean_step_time 1436.27, it/s 0.7
  5086. [0130 17:54:59 @multigpu.py:323] [p0115]  step: count(1025), step_time 1415.06, mean_step_time 1435.87, it/s 0.7
  5087. [0130 17:54:59 @multigpu.py:323] [p0576]  step: count(1049), step_time 1498.11, mean_step_time 1436.52, it/s 0.7
  5088. [0130 17:55:00 @multigpu.py:323] [p0574]  step: count(1041), step_time 1429.55, mean_step_time 1434.22, it/s 0.7
  5089. [0130 17:55:01 @multigpu.py:323] [p0576]  step: count(1050), step_time 1387.85, mean_step_time 1433.92, it/s 0.7
  5090. [0130 17:55:01 @multigpu.py:323] [p0115]  step: count(1026), step_time 1455.92, mean_step_time 1437.49, it/s 0.7
  5091. [0130 17:55:02 @multigpu.py:323] [p0574]  step: count(1042), step_time 1404.44, mean_step_time 1434.0, it/s 0.7
  5092. [0130 17:55:02 @multigpu.py:323] [p0576]  step: count(1051), step_time 1454.33, mean_step_time 1434.37, it/s 0.7
  5093. [0130 17:55:02 @multigpu.py:323] [p0115]  step: count(1027), step_time 1447.54, mean_step_time 1438.77, it/s 0.7
  5094. sending to address tcp://p0112:61216
  5095. ##### Sending to neptune:  online_score :  0.417075758311 , 1.6 #####
  5096. [u'online', 1.6]
  5097. ##### Sending to neptune:  active_workers :  0.41707579467 , 3 #####
  5098. receiving
  5099. [0130 17:55:04 @multigpu.py:323] [p0115]  step: count(1028), step_time 1484.34, mean_step_time 1440.64, it/s 0.69
  5100. [0130 17:55:04 @multigpu.py:323] [p0576]  step: count(1052), step_time 1519.52, mean_step_time 1440.58, it/s 0.69
  5101. [0130 17:55:04 @multigpu.py:323] [p0574]  step: count(1043), step_time 1747.43, mean_step_time 1440.63, it/s 0.69
  5102. [0130 17:55:05 @multigpu.py:323] [p0576]  step: count(1053), step_time 1418.04, mean_step_time 1440.68, it/s 0.69
  5103. [0130 17:55:05 @multigpu.py:323] [p0574]  step: count(1044), step_time 1417.8, mean_step_time 1442.0, it/s 0.69
  5104. [0130 17:55:05 @multigpu.py:323] [p0115]  step: count(1029), step_time 1472.79, mean_step_time 1442.52, it/s 0.69
  5105. [0130 17:55:06 @multigpu.py:323] [p0574]  step: count(1045), step_time 1394.07, mean_step_time 1438.98, it/s 0.69
  5106. [0130 17:55:06 @multigpu.py:323] [p0576]  step: count(1054), step_time 1410.82, mean_step_time 1438.52, it/s 0.7
  5107. [0130 17:55:06 @multigpu.py:323] [p0115]  step: count(1030), step_time 1420.81, mean_step_time 1443.74, it/s 0.69
  5108. [0130 17:55:08 @multigpu.py:323] [p0574]  step: count(1046), step_time 1393.73, mean_step_time 1435.46, it/s 0.7
  5109. [0130 17:55:08 @multigpu.py:323] [p0576]  step: count(1055), step_time 1458.42, mean_step_time 1437.25, it/s 0.7
  5110. [0130 17:55:08 @multigpu.py:323] [p0115]  step: count(1031), step_time 1485.93, mean_step_time 1445.16, it/s 0.69
  5111. [0130 17:55:09 @multigpu.py:323] [p0574]  step: count(1047), step_time 1478.91, mean_step_time 1437.34, it/s 0.7
  5112. [0130 17:55:09 @multigpu.py:323] [p0576]  step: count(1056), step_time 1488.08, mean_step_time 1440.48, it/s 0.69
  5113. [0130 17:55:09 @multigpu.py:323] [p0115]  step: count(1032), step_time 1414.59, mean_step_time 1443.06, it/s 0.69
  5114. sending to address tcp://p0112:61216
  5115. ##### Sending to neptune:  online_score :  0.419183745517 , 1.7 #####
  5116. [u'online', 1.7]
  5117. receiving
  5118. [0130 17:55:11 @multigpu.py:323] [p0574]  step: count(1048), step_time 1408.46, mean_step_time 1436.16, it/s 0.7
  5119. [0130 17:55:11 @multigpu.py:323] [p0576]  step: count(1057), step_time 1376.61, mean_step_time 1438.7, it/s 0.7
  5120. [0130 17:55:11 @multigpu.py:323] [p0115]  step: count(1033), step_time 1422.66, mean_step_time 1442.19, it/s 0.69
  5121. [0130 17:55:12 @multigpu.py:323] [p0574]  step: count(1049), step_time 1398.56, mean_step_time 1433.16, it/s 0.7
  5122. [0130 17:55:12 @multigpu.py:323] [p0576]  step: count(1058), step_time 1436.46, mean_step_time 1438.52, it/s 0.7
  5123. [0130 17:55:12 @multigpu.py:323] [p0115]  step: count(1034), step_time 1444.23, mean_step_time 1443.26, it/s 0.69
  5124. [0130 17:55:13 @multigpu.py:323] [p0574]  step: count(1050), step_time 1419.26, mean_step_time 1432.29, it/s 0.7
  5125. [0130 17:55:14 @multigpu.py:323] [p0576]  step: count(1059), step_time 1464.07, mean_step_time 1442.54, it/s 0.69
  5126. [0130 17:55:14 @multigpu.py:323] [p0115]  step: count(1035), step_time 1432.57, mean_step_time 1442.88, it/s 0.69
  5127. [0130 17:55:15 @multigpu.py:323] [p0574]  step: count(1051), step_time 1429.73, mean_step_time 1431.13, it/s 0.7
  5128. [0130 17:55:15 @multigpu.py:323] [p0576]  step: count(1060), step_time 1391.39, mean_step_time 1441.11, it/s 0.69
  5129. [0130 17:55:15 @multigpu.py:323] [p0115]  step: count(1036), step_time 1421.45, mean_step_time 1438.93, it/s 0.69
  5130. [0130 17:55:16 @multigpu.py:323] [p0574]  step: count(1052), step_time 1391.14, mean_step_time 1430.65, it/s 0.7
  5131. [0130 17:55:16 @multigpu.py:323] [p0576]  step: count(1061), step_time 1390.34, mean_step_time 1439.87, it/s 0.69
  5132. [0130 17:55:16 @multigpu.py:323] [p0115]  step: count(1037), step_time 1424.11, mean_step_time 1439.48, it/s 0.69
  5133. sending to address tcp://p0112:61216
  5134. ##### Sending to neptune:  online_score :  0.421175178554 , 0.9 #####
  5135. [u'online', 0.9]
  5136. receiving
  5137. [0130 17:55:18 @multigpu.py:323] [p0574]  step: count(1053), step_time 1493.1, mean_step_time 1432.63, it/s 0.7
  5138. [0130 17:55:18 @multigpu.py:323] [p0576]  step: count(1062), step_time 1423.26, mean_step_time 1442.71, it/s 0.69
  5139. [0130 17:55:18 @multigpu.py:323] [p0115]  step: count(1038), step_time 1529.8, mean_step_time 1444.2, it/s 0.69
  5140. [0130 17:55:19 @multigpu.py:323] [p0574]  step: count(1054), step_time 1428.93, mean_step_time 1434.13, it/s 0.7
  5141. [0130 17:55:19 @multigpu.py:323] [p0576]  step: count(1063), step_time 1395.72, mean_step_time 1438.78, it/s 0.7
  5142. [0130 17:55:19 @multigpu.py:323] [p0115]  step: count(1039), step_time 1404.79, mean_step_time 1443.11, it/s 0.69
  5143. [0130 17:55:21 @multigpu.py:323] [p0576]  step: count(1064), step_time 1376.76, mean_step_time 1433.59, it/s 0.7
  5144. [0130 17:55:21 @multigpu.py:323] [p0574]  step: count(1055), step_time 1439.54, mean_step_time 1435.19, it/s 0.7
  5145. [0130 17:55:21 @multigpu.py:323] [p0115]  step: count(1040), step_time 1448.15, mean_step_time 1444.78, it/s 0.69
  5146. [0130 17:55:22 @multigpu.py:323] [p0576]  step: count(1065), step_time 1459.01, mean_step_time 1431.51, it/s 0.7
  5147. [0130 17:55:22 @multigpu.py:323] [p0574]  step: count(1056), step_time 1463.82, mean_step_time 1439.71, it/s 0.69
  5148. [0130 17:55:22 @multigpu.py:323] [p0115]  step: count(1041), step_time 1461.35, mean_step_time 1445.5, it/s 0.69
  5149. [0130 17:55:23 @multigpu.py:323] [p0576]  step: count(1066), step_time 1372.05, mean_step_time 1430.91, it/s 0.7
  5150. [0130 17:55:24 @multigpu.py:323] [p0574]  step: count(1057), step_time 1457.14, mean_step_time 1440.27, it/s 0.69
  5151. [0130 17:55:24 @multigpu.py:323] [p0115]  step: count(1042), step_time 1423.25, mean_step_time 1444.08, it/s 0.69
  5152. [0130 17:55:25 @multigpu.py:323] [p0576]  step: count(1067), step_time 1508.45, mean_step_time 1432.4, it/s 0.7
  5153. [0130 17:55:25 @multigpu.py:323] [p0574]  step: count(1058), step_time 1400.48, mean_step_time 1440.97, it/s 0.69
  5154. [0130 17:55:25 @multigpu.py:323] [p0115]  step: count(1043), step_time 1475.45, mean_step_time 1446.09, it/s 0.69
  5155. [0130 17:55:26 @multigpu.py:323] [p0576]  step: count(1068), step_time 1395.59, mean_step_time 1431.24, it/s 0.7
  5156. [0130 17:55:26 @multigpu.py:323] [p0574]  step: count(1059), step_time 1448.04, mean_step_time 1441.98, it/s 0.69
  5157. [0130 17:55:27 @multigpu.py:323] [p0115]  step: count(1044), step_time 1390.38, mean_step_time 1443.76, it/s 0.69
  5158. [0130 17:55:28 @multigpu.py:323] [p0576]  step: count(1069), step_time 1399.62, mean_step_time 1426.32, it/s 0.7
  5159. [0130 17:55:28 @multigpu.py:323] [p0574]  step: count(1060), step_time 1422.05, mean_step_time 1443.31, it/s 0.69
  5160. [0130 17:55:28 @multigpu.py:323] [p0115]  step: count(1045), step_time 1472.17, mean_step_time 1446.62, it/s 0.69
  5161. [0130 17:55:29 @multigpu.py:323] [p0576]  step: count(1070), step_time 1414.84, mean_step_time 1427.67, it/s 0.7
  5162. [0130 17:55:29 @multigpu.py:323] [p0574]  step: count(1061), step_time 1457.26, mean_step_time 1444.69, it/s 0.69
  5163. [0130 17:55:30 @multigpu.py:323] [p0115]  step: count(1046), step_time 1464.66, mean_step_time 1447.05, it/s 0.69
  5164. [0130 17:55:31 @multigpu.py:323] [p0576]  step: count(1071), step_time 1431.19, mean_step_time 1426.51, it/s 0.7
  5165. [0130 17:55:31 @multigpu.py:323] [p0574]  step: count(1062), step_time 1417.34, mean_step_time 1445.34, it/s 0.69
  5166. [0130 17:55:31 @multigpu.py:323] [p0115]  step: count(1047), step_time 1442.89, mean_step_time 1446.82, it/s 0.69
  5167. sending to address tcp://p0112:61216
  5168. ##### Sending to neptune:  online_score :  0.425126151906 , 1.7 #####
  5169. [u'online', 1.7]
  5170. receiving
  5171. sending to address tcp://p0112:61216
  5172. ##### Sending to neptune:  online_score :  0.42519397027 , 1.3 #####
  5173. [u'online', 1.3]
  5174. receiving
  5175. [0130 17:55:32 @multigpu.py:323] [p0574]  step: count(1063), step_time 1743.95, mean_step_time 1445.16, it/s 0.69
  5176. [0130 17:55:32 @multigpu.py:323] [p0576]  step: count(1072), step_time 1891.85, mean_step_time 1445.13, it/s 0.69
  5177. [0130 17:55:32 @multigpu.py:323] [p0115]  step: count(1048), step_time 1451.52, mean_step_time 1445.18, it/s 0.69
  5178. [0130 17:55:34 @multigpu.py:323] [p0576]  step: count(1073), step_time 1409.99, mean_step_time 1444.73, it/s 0.69
  5179. [0130 17:55:34 @multigpu.py:323] [p0574]  step: count(1064), step_time 1415.11, mean_step_time 1445.03, it/s 0.69
  5180. [0130 17:55:34 @multigpu.py:323] [p0115]  step: count(1049), step_time 1455.97, mean_step_time 1444.34, it/s 0.69
  5181. [0130 17:55:35 @multigpu.py:323] [p0576]  step: count(1074), step_time 1400.35, mean_step_time 1444.2, it/s 0.69
  5182. [0130 17:55:35 @multigpu.py:323] [p0574]  step: count(1065), step_time 1401.07, mean_step_time 1445.38, it/s 0.69
  5183. [0130 17:55:35 @multigpu.py:323] [p0115]  step: count(1050), step_time 1371.89, mean_step_time 1441.89, it/s 0.69
  5184. [0130 17:55:37 @multigpu.py:323] [p0574]  step: count(1066), step_time 1459.04, mean_step_time 1448.65, it/s 0.69
  5185. [0130 17:55:37 @multigpu.py:323] [p0115]  step: count(1051), step_time 1458.72, mean_step_time 1440.53, it/s 0.69
  5186. [0130 17:55:37 @multigpu.py:323] [p0576]  step: count(1075), step_time 1482.75, mean_step_time 1445.42, it/s 0.69
  5187. [0130 17:55:38 @multigpu.py:323] [p0574]  step: count(1067), step_time 1401.78, mean_step_time 1444.79, it/s 0.69
  5188. [0130 17:55:38 @multigpu.py:323] [p0115]  step: count(1052), step_time 1408.94, mean_step_time 1440.25, it/s 0.69
  5189. [0130 17:55:38 @multigpu.py:323] [p0576]  step: count(1076), step_time 1422.88, mean_step_time 1442.16, it/s 0.69
  5190. [0130 17:55:40 @multigpu.py:323] [p0576]  step: count(1077), step_time 1375.78, mean_step_time 1442.12, it/s 0.69
  5191. [0130 17:55:40 @multigpu.py:323] [p0574]  step: count(1068), step_time 1455.3, mean_step_time 1447.13, it/s 0.69
  5192. [0130 17:55:40 @multigpu.py:323] [p0115]  step: count(1053), step_time 1468.53, mean_step_time 1442.54, it/s 0.69
  5193. [0130 17:55:41 @multigpu.py:323] [p0574]  step: count(1069), step_time 1394.51, mean_step_time 1446.93, it/s 0.69
  5194. [0130 17:55:41 @multigpu.py:323] [p0576]  step: count(1078), step_time 1477.76, mean_step_time 1444.18, it/s 0.69
  5195. [0130 17:55:41 @multigpu.py:323] [p0115]  step: count(1054), step_time 1483.98, mean_step_time 1444.53, it/s 0.69
  5196. [0130 17:55:42 @multigpu.py:323] [p0574]  step: count(1070), step_time 1435.69, mean_step_time 1447.75, it/s 0.69
  5197. [0130 17:55:42 @multigpu.py:323] [p0576]  step: count(1079), step_time 1457.34, mean_step_time 1443.85, it/s 0.69
  5198. [0130 17:55:43 @multigpu.py:323] [p0115]  step: count(1055), step_time 1441.54, mean_step_time 1444.98, it/s 0.69
  5199. [0130 17:55:44 @multigpu.py:323] [p0574]  step: count(1071), step_time 1401.82, mean_step_time 1446.35, it/s 0.69
  5200. [0130 17:55:44 @multigpu.py:323] [p0576]  step: count(1080), step_time 1395.95, mean_step_time 1444.07, it/s 0.69
  5201. [0130 17:55:44 @multigpu.py:323] [p0115]  step: count(1056), step_time 1444.04, mean_step_time 1446.11, it/s 0.69
  5202. [0130 17:55:45 @multigpu.py:323] [p0574]  step: count(1072), step_time 1388.22, mean_step_time 1446.21, it/s 0.69
  5203. [0130 17:55:45 @multigpu.py:323] [p0576]  step: count(1081), step_time 1439.41, mean_step_time 1446.53, it/s 0.69
  5204. [0130 17:55:45 @multigpu.py:323] [p0115]  step: count(1057), step_time 1424.31, mean_step_time 1446.12, it/s 0.69
  5205. [0130 17:55:47 @multigpu.py:323] [p0574]  step: count(1073), step_time 1442.95, mean_step_time 1443.7, it/s 0.69
  5206. [0130 17:55:47 @multigpu.py:323] [p0576]  step: count(1082), step_time 1515.43, mean_step_time 1451.14, it/s 0.69
  5207. [0130 17:55:47 @multigpu.py:323] [p0115]  step: count(1058), step_time 1443.79, mean_step_time 1441.82, it/s 0.69
  5208. [0130 17:55:48 @multigpu.py:323] [p0574]  step: count(1074), step_time 1441.68, mean_step_time 1444.34, it/s 0.69
  5209. [0130 17:55:48 @multigpu.py:323] [p0576]  step: count(1083), step_time 1382.39, mean_step_time 1450.47, it/s 0.69
  5210. [0130 17:55:48 @multigpu.py:323] [p0115]  step: count(1059), step_time 1439.56, mean_step_time 1443.56, it/s 0.69
  5211. [0130 17:55:49 @multigpu.py:323] [p0574]  step: count(1075), step_time 1374.67, mean_step_time 1441.1, it/s 0.69
  5212. [0130 17:55:50 @multigpu.py:323] [p0576]  step: count(1084), step_time 1449.83, mean_step_time 1454.12, it/s 0.69
  5213. [0130 17:55:50 @multigpu.py:323] [p0115]  step: count(1060), step_time 1461.11, mean_step_time 1444.2, it/s 0.69
  5214. [0130 17:55:51 @multigpu.py:323] [p0574]  step: count(1076), step_time 1471.64, mean_step_time 1441.49, it/s 0.69
  5215. [0130 17:55:51 @multigpu.py:323] [p0576]  step: count(1085), step_time 1367.67, mean_step_time 1449.56, it/s 0.69
  5216. [0130 17:55:51 @multigpu.py:323] [p0115]  step: count(1061), step_time 1464.32, mean_step_time 1444.35, it/s 0.69
  5217. [0130 17:55:52 @multigpu.py:323] [p0574]  step: count(1077), step_time 1408.18, mean_step_time 1439.04, it/s 0.69
  5218. [0130 17:55:53 @multigpu.py:323] [p0576]  step: count(1086), step_time 1499.73, mean_step_time 1455.94, it/s 0.69
  5219. [0130 17:55:53 @multigpu.py:323] [p0115]  step: count(1062), step_time 1418.45, mean_step_time 1444.11, it/s 0.69
  5220. sending to address tcp://p0112:61216
  5221. ##### Sending to neptune:  online_score :  0.431302230491 , 1.3 #####
  5222. [u'online', 1.3]
  5223. receiving
  5224. [0130 17:55:54 @multigpu.py:323] [p0574]  step: count(1078), step_time 1357.14, mean_step_time 1436.87, it/s 0.7
  5225. [0130 17:55:54 @multigpu.py:323] [p0576]  step: count(1087), step_time 1445.39, mean_step_time 1452.79, it/s 0.69
  5226. [0130 17:55:54 @multigpu.py:323] [p0115]  step: count(1063), step_time 1366.45, mean_step_time 1438.66, it/s 0.7
  5227. [0130 17:55:55 @multigpu.py:323] [p0574]  step: count(1079), step_time 1467.04, mean_step_time 1437.82, it/s 0.7
  5228. [0130 17:55:55 @multigpu.py:323] [p0576]  step: count(1088), step_time 1401.09, mean_step_time 1453.06, it/s 0.69
  5229. [0130 17:55:55 @multigpu.py:323] [p0115]  step: count(1064), step_time 1437.42, mean_step_time 1441.01, it/s 0.69
  5230. [0130 17:55:57 @multigpu.py:323] [p0574]  step: count(1080), step_time 1421.0, mean_step_time 1437.77, it/s 0.7
  5231. [0130 17:55:57 @multigpu.py:323] [p0576]  step: count(1089), step_time 1454.31, mean_step_time 1455.8, it/s 0.69
  5232. [0130 17:55:57 @multigpu.py:323] [p0115]  step: count(1065), step_time 1415.52, mean_step_time 1438.18, it/s 0.7
  5233. [0130 17:55:58 @multigpu.py:323] [p0574]  step: count(1081), step_time 1413.1, mean_step_time 1435.56, it/s 0.7
  5234. [0130 17:55:58 @multigpu.py:323] [p0576]  step: count(1090), step_time 1415.9, mean_step_time 1455.85, it/s 0.69
  5235. [0130 17:55:58 @multigpu.py:323] [p0115]  step: count(1066), step_time 1426.7, mean_step_time 1436.28, it/s 0.7
  5236. [0130 17:55:59 @multigpu.py:323] [p0574]  step: count(1082), step_time 1393.46, mean_step_time 1434.37, it/s 0.7
  5237. [0130 17:56:00 @multigpu.py:323] [p0576]  step: count(1091), step_time 1428.45, mean_step_time 1455.71, it/s 0.69
  5238. [0130 17:56:00 @multigpu.py:323] [p0115]  step: count(1067), step_time 1429.34, mean_step_time 1435.61, it/s 0.7
  5239. [0130 17:56:01 @multigpu.py:323] [p0115]  step: count(1068), step_time 1448.33, mean_step_time 1435.45, it/s 0.7
  5240. [0130 17:56:01 @multigpu.py:323] [p0576]  step: count(1092), step_time 1486.72, mean_step_time 1435.46, it/s 0.7
  5241. [0130 17:56:01 @multigpu.py:323] [p0574]  step: count(1083), step_time 1768.75, mean_step_time 1435.61, it/s 0.7
  5242. [0130 17:56:03 @multigpu.py:323] [p0576]  step: count(1093), step_time 1380.93, mean_step_time 1434.0, it/s 0.7
  5243. [0130 17:56:03 @multigpu.py:323] [p0115]  step: count(1069), step_time 1439.15, mean_step_time 1434.6, it/s 0.7
  5244. [0130 17:56:03 @multigpu.py:323] [p0574]  step: count(1084), step_time 1507.57, mean_step_time 1440.23, it/s 0.69
  5245. sending to address tcp://p0112:61216
  5246. ##### Sending to neptune:  online_score :  0.433990958863 , 1.7 #####
  5247. [u'online', 1.7]
  5248. ##### Sending to neptune:  active_workers :  0.433991316888 , 3 #####
  5249. receiving
  5250. [0130 17:56:04 @multigpu.py:323] [p0576]  step: count(1094), step_time 1471.82, mean_step_time 1437.58, it/s 0.7
  5251. [0130 17:56:04 @multigpu.py:323] [p0115]  step: count(1070), step_time 1424.13, mean_step_time 1437.22, it/s 0.7
  5252. [0130 17:56:04 @multigpu.py:323] [p0574]  step: count(1085), step_time 1368.88, mean_step_time 1438.62, it/s 0.7
  5253. sending to address tcp://p0112:61216
  5254. ##### Sending to neptune:  online_score :  0.434331390262 , 1.6 #####
  5255. [u'online', 1.6]
  5256. receiving
  5257. [0130 17:56:05 @multigpu.py:323] [p0576]  step: count(1095), step_time 1405.61, mean_step_time 1433.72, it/s 0.7
  5258. [0130 17:56:05 @multigpu.py:323] [p0115]  step: count(1071), step_time 1422.46, mean_step_time 1435.4, it/s 0.7
  5259. [0130 17:56:05 @multigpu.py:323] [p0574]  step: count(1086), step_time 1438.94, mean_step_time 1437.62, it/s 0.7
  5260. [0130 17:56:07 @multigpu.py:323] [p0115]  step: count(1072), step_time 1412.86, mean_step_time 1435.6, it/s 0.7
  5261. [0130 17:56:07 @multigpu.py:323] [p0576]  step: count(1096), step_time 1456.46, mean_step_time 1435.4, it/s 0.7
  5262. [0130 17:56:07 @multigpu.py:323] [p0574]  step: count(1087), step_time 1446.37, mean_step_time 1439.85, it/s 0.69
  5263. [0130 17:56:08 @multigpu.py:323] [p0576]  step: count(1097), step_time 1388.13, mean_step_time 1436.02, it/s 0.7
  5264. [0130 17:56:08 @multigpu.py:323] [p0115]  step: count(1073), step_time 1441.22, mean_step_time 1434.23, it/s 0.7
  5265. [0130 17:56:08 @multigpu.py:323] [p0574]  step: count(1088), step_time 1420.67, mean_step_time 1438.11, it/s 0.7
  5266. [0130 17:56:10 @multigpu.py:323] [p0576]  step: count(1098), step_time 1390.21, mean_step_time 1431.64, it/s 0.7
  5267. [0130 17:56:10 @multigpu.py:323] [p0115]  step: count(1074), step_time 1403.55, mean_step_time 1430.21, it/s 0.7
  5268. [0130 17:56:10 @multigpu.py:323] [p0574]  step: count(1089), step_time 1454.05, mean_step_time 1441.09, it/s 0.69
  5269. [0130 17:56:11 @multigpu.py:323] [p0576]  step: count(1099), step_time 1401.55, mean_step_time 1428.85, it/s 0.7
  5270. [0130 17:56:11 @multigpu.py:323] [p0115]  step: count(1075), step_time 1449.39, mean_step_time 1430.6, it/s 0.7
  5271. [0130 17:56:11 @multigpu.py:323] [p0574]  step: count(1090), step_time 1442.23, mean_step_time 1441.42, it/s 0.69
  5272. [0130 17:56:13 @multigpu.py:323] [p0576]  step: count(1100), step_time 1444.12, mean_step_time 1431.26, it/s 0.7
  5273. sending debugging info...
  5274. sending to address tcp://p0112:61216
  5275. ##### Sending to neptune:  mean_delay :  0.436645072765 , 0.0 #####
  5276. sending to address tcp://p0112:61216
  5277. ##### Sending to neptune:  max_delay :  0.436645072765 , -0.0 #####
  5278. sending to address tcp://p0112:61216
  5279. ##### Sending to neptune:  min_delay :  0.436645072765 , -0.0 #####
  5280. [u'delays', [0.0, -0.0, -0.0]]
  5281. receiving
  5282. ##### Sending to neptune:  cost :  0.436646228035 , -0.0118187069893 #####
  5283. ##### Sending to neptune:  policy_loss :  0.436646228035 , -0.450571477413 #####
  5284. ##### Sending to neptune:  xentropy_loss :  0.436646228035 , -2.28830933571 #####
  5285. ##### Sending to neptune:  value_loss :  0.436646228035 , 1.22608649731 #####
  5286. ##### Sending to neptune:  advantage :  0.436646228035 , 0.00196923920885 #####
  5287. ##### Sending to neptune:  pred_reward :  0.436646228035 , 0.406696081161 #####
  5288. ##### Sending to neptune:  max_logit :  0.436646228035 , 0.194596156478 #####
  5289. [u'loss', -0.01181870698928833, -0.4505714774131775, -2.288309335708618, 1.2260864973068237, 0.0019692392088472843, 0.406696081161499, 0.19459615647792816]
  5290. receiving
  5291. ##### Sending to neptune:  active_relus :  0.436649463574 , 9426596.31 #####
  5292. ##### Sending to neptune:  dp_per_s :  0.436649463574 , 89.1332720899 #####
  5293. [u'other', 9426596.31, 89.13327208992234]
  5294. receiving
  5295. [0130 17:56:13 @multigpu.py:323] [p0115]  step: count(1076), step_time 1434.04, mean_step_time 1430.1, it/s 0.7
  5296. [0130 17:56:13 @multigpu.py:323] [p0574]  step: count(1091), step_time 1448.36, mean_step_time 1443.75, it/s 0.69
  5297. sending to address tcp://p0112:61216
  5298. ##### Sending to neptune:  online_score :  0.436772664388 , 1.7 #####
  5299. [u'online', 1.7]
  5300. receiving
  5301. [0130 17:56:14 @multigpu.py:323] [p0576]  step: count(1101), step_time 1425.92, mean_step_time 1430.58, it/s 0.7
  5302. [0130 17:56:14 @multigpu.py:323] [p0115]  step: count(1077), step_time 1443.31, mean_step_time 1431.05, it/s 0.7
  5303. [0130 17:56:14 @multigpu.py:323] [p0574]  step: count(1092), step_time 1399.08, mean_step_time 1444.29, it/s 0.69
  5304. [0130 17:56:15 @multigpu.py:323] [p0576]  step: count(1102), step_time 1399.16, mean_step_time 1424.77, it/s 0.7
  5305. [0130 17:56:15 @multigpu.py:323] [p0115]  step: count(1078), step_time 1415.94, mean_step_time 1429.66, it/s 0.7
  5306. [0130 17:56:16 @multigpu.py:323] [p0574]  step: count(1093), step_time 1449.26, mean_step_time 1444.6, it/s 0.69
  5307. [0130 17:56:17 @multigpu.py:323] [p0576]  step: count(1103), step_time 1461.29, mean_step_time 1428.71, it/s 0.7
  5308. [0130 17:56:17 @multigpu.py:323] [p0115]  step: count(1079), step_time 1432.04, mean_step_time 1429.29, it/s 0.7
  5309. [0130 17:56:17 @multigpu.py:323] [p0574]  step: count(1094), step_time 1396.09, mean_step_time 1442.32, it/s 0.69
  5310. [0130 17:56:18 @multigpu.py:323] [p0576]  step: count(1104), step_time 1415.22, mean_step_time 1426.98, it/s 0.7
  5311. [0130 17:56:18 @multigpu.py:323] [p0115]  step: count(1080), step_time 1432.88, mean_step_time 1427.87, it/s 0.7
  5312. [0130 17:56:18 @multigpu.py:323] [p0574]  step: count(1095), step_time 1397.45, mean_step_time 1443.46, it/s 0.69
  5313. [0130 17:56:20 @multigpu.py:323] [p0576]  step: count(1105), step_time 1417.93, mean_step_time 1429.5, it/s 0.7
  5314. [0130 17:56:20 @multigpu.py:323] [p0115]  step: count(1081), step_time 1441.1, mean_step_time 1426.71, it/s 0.7
  5315. [0130 17:56:20 @multigpu.py:323] [p0574]  step: count(1096), step_time 1459.25, mean_step_time 1442.84, it/s 0.69
  5316. [0130 17:56:21 @multigpu.py:323] [p0576]  step: count(1106), step_time 1440.26, mean_step_time 1426.52, it/s 0.7
  5317. [0130 17:56:21 @multigpu.py:323] [p0115]  step: count(1082), step_time 1419.84, mean_step_time 1426.78, it/s 0.7
  5318. [0130 17:56:21 @multigpu.py:323] [p0574]  step: count(1097), step_time 1449.18, mean_step_time 1444.89, it/s 0.69
  5319. [0130 17:56:23 @multigpu.py:323] [p0576]  step: count(1107), step_time 1465.27, mean_step_time 1427.52, it/s 0.7
  5320. [0130 17:56:23 @multigpu.py:323] [p0115]  step: count(1083), step_time 1423.24, mean_step_time 1429.62, it/s 0.7
  5321. [0130 17:56:23 @multigpu.py:323] [p0574]  step: count(1098), step_time 1523.52, mean_step_time 1453.21, it/s 0.69
  5322. [0130 17:56:24 @multigpu.py:323] [p0576]  step: count(1108), step_time 1403.19, mean_step_time 1427.62, it/s 0.7
  5323. [0130 17:56:24 @multigpu.py:323] [p0115]  step: count(1084), step_time 1477.21, mean_step_time 1431.61, it/s 0.7
  5324. [0130 17:56:24 @multigpu.py:323] [p0574]  step: count(1099), step_time 1472.36, mean_step_time 1453.48, it/s 0.69
  5325. [0130 17:56:25 @multigpu.py:323] [p0576]  step: count(1109), step_time 1415.13, mean_step_time 1425.66, it/s 0.7
  5326. [0130 17:56:25 @multigpu.py:323] [p0115]  step: count(1085), step_time 1391.89, mean_step_time 1430.43, it/s 0.7
  5327. [0130 17:56:26 @multigpu.py:323] [p0574]  step: count(1100), step_time 1408.9, mean_step_time 1452.87, it/s 0.69
  5328. sending debugging info...
  5329. sending to address tcp://p0112:61216
  5330. ##### Sending to neptune:  mean_delay :  0.440297738579 , 0.0 #####
  5331. ##### Sending to neptune:  max_delay :  0.440297738579 , -0.0 #####
  5332. sending to address tcp://p0112:61216
  5333. ##### Sending to neptune:  min_delay :  0.440297738579 , -0.0 #####
  5334. [u'delays', [0.0, -0.0, -0.0]]
  5335. receiving
  5336. ##### Sending to neptune:  cost :  0.440298264954 , -0.00412045605481 #####
  5337. ##### Sending to neptune:  policy_loss :  0.440298264954 , 0.338887602091 #####
  5338. sending to address tcp://p0112:61216
  5339. ##### Sending to neptune:  xentropy_loss :  0.440298264954 , -2.28836631775 #####
  5340. ##### Sending to neptune:  value_loss :  0.440298264954 , 1.42206013203 #####
  5341. ##### Sending to neptune:  advantage :  0.440298264954 , -0.00142946292181 #####
  5342. ##### Sending to neptune:  pred_reward :  0.440298264954 , 0.410076230764 #####
  5343. ##### Sending to neptune:  max_logit :  0.440298264954 , 0.194619998336 #####
  5344. [u'loss', -0.004120456054806709, 0.33888760209083557, -2.2883663177490234, 1.4220601320266724, -0.0014294629218056798, 0.41007623076438904, 0.19461999833583832]
  5345. receiving
  5346. ##### Sending to neptune:  active_relus :  0.440298698876 , 9435936.15 #####
  5347. ##### Sending to neptune:  dp_per_s :  0.440298698876 , 89.0701042424 #####
  5348. [u'other', 9435936.15, 89.07010424237725]
  5349. receiving
  5350. [0130 17:56:27 @multigpu.py:323] [p0576]  step: count(1110), step_time 1438.01, mean_step_time 1426.77, it/s 0.7
  5351. [0130 17:56:27 @multigpu.py:323] [p0115]  step: count(1086), step_time 1432.05, mean_step_time 1430.7, it/s 0.7
  5352. [0130 17:56:27 @multigpu.py:323] [p0574]  step: count(1101), step_time 1443.19, mean_step_time 1454.38, it/s 0.69
  5353. [0130 17:56:28 @multigpu.py:323] [p0576]  step: count(1111), step_time 1431.44, mean_step_time 1426.92, it/s 0.7
  5354. [0130 17:56:28 @multigpu.py:323] [p0115]  step: count(1087), step_time 1435.4, mean_step_time 1431.0, it/s 0.7
  5355. [0130 17:56:29 @multigpu.py:323] [p0574]  step: count(1102), step_time 1433.72, mean_step_time 1456.39, it/s 0.69
  5356. sending to address tcp://p0112:61216
  5357. ##### Sending to neptune:  online_score :  0.441290416651 , 1.3 #####
  5358. [u'online', 1.3]
  5359. receiving
  5360. [0130 17:56:30 @multigpu.py:323] [p0115]  step: count(1088), step_time 1701.67, mean_step_time 1443.67, it/s 0.69
  5361. [0130 17:56:30 @multigpu.py:323] [p0574]  step: count(1103), step_time 1512.89, mean_step_time 1443.6, it/s 0.69
  5362. [0130 17:56:30 @multigpu.py:323] [p0576]  step: count(1112), step_time 1822.69, mean_step_time 1443.72, it/s 0.69
  5363. [0130 17:56:32 @multigpu.py:323] [p0115]  step: count(1089), step_time 1452.35, mean_step_time 1444.33, it/s 0.69
  5364. [0130 17:56:32 @multigpu.py:323] [p0576]  step: count(1113), step_time 1466.74, mean_step_time 1448.01, it/s 0.69
  5365. [0130 17:56:32 @multigpu.py:323] [p0574]  step: count(1104), step_time 1532.74, mean_step_time 1444.86, it/s 0.69
  5366. [0130 17:56:33 @multigpu.py:323] [p0115]  step: count(1090), step_time 1384.42, mean_step_time 1442.34, it/s 0.69
  5367. [0130 17:56:33 @multigpu.py:323] [p0576]  step: count(1114), step_time 1431.12, mean_step_time 1445.97, it/s 0.69
  5368. [0130 17:56:33 @multigpu.py:323] [p0574]  step: count(1105), step_time 1461.45, mean_step_time 1449.49, it/s 0.69
  5369. sending to address tcp://p0112:61216
  5370. ##### Sending to neptune:  online_score :  0.442541402446 , 1.8 #####
  5371. [u'online', 1.8]
  5372. receiving
  5373. [0130 17:56:34 @multigpu.py:323] [p0115]  step: count(1091), step_time 1429.84, mean_step_time 1442.71, it/s 0.69
  5374. [0130 17:56:34 @multigpu.py:323] [p0576]  step: count(1115), step_time 1403.25, mean_step_time 1445.86, it/s 0.69
  5375. [0130 17:56:35 @multigpu.py:323] [p0574]  step: count(1106), step_time 1487.84, mean_step_time 1451.93, it/s 0.69
  5376. [0130 17:56:36 @multigpu.py:323] [p0115]  step: count(1092), step_time 1448.2, mean_step_time 1444.48, it/s 0.69
  5377. [0130 17:56:36 @multigpu.py:323] [p0576]  step: count(1116), step_time 1437.0, mean_step_time 1444.88, it/s 0.69
  5378. [0130 17:56:36 @multigpu.py:323] [p0574]  step: count(1107), step_time 1386.29, mean_step_time 1448.93, it/s 0.69
  5379. [0130 17:56:37 @multigpu.py:323] [p0115]  step: count(1093), step_time 1401.62, mean_step_time 1442.5, it/s 0.69
  5380. [0130 17:56:37 @multigpu.py:323] [p0576]  step: count(1117), step_time 1425.33, mean_step_time 1446.74, it/s 0.69
  5381. [0130 17:56:37 @multigpu.py:323] [p0574]  step: count(1108), step_time 1515.78, mean_step_time 1453.68, it/s 0.69
  5382. [0130 17:56:39 @multigpu.py:323] [p0115]  step: count(1094), step_time 1457.24, mean_step_time 1445.18, it/s 0.69
  5383. [0130 17:56:39 @multigpu.py:323] [p0576]  step: count(1118), step_time 1448.3, mean_step_time 1449.65, it/s 0.69
  5384. [0130 17:56:39 @multigpu.py:323] [p0574]  step: count(1109), step_time 1421.13, mean_step_time 1452.04, it/s 0.69
  5385. [0130 17:56:40 @multigpu.py:323] [p0115]  step: count(1095), step_time 1438.18, mean_step_time 1444.62, it/s 0.69
  5386. [0130 17:56:40 @multigpu.py:323] [p0576]  step: count(1119), step_time 1476.4, mean_step_time 1453.39, it/s 0.69
  5387. sending to address tcp://p0112:61216
  5388. ##### Sending to neptune:  online_score :  0.444324848586 , 0.9 #####
  5389. [u'online', 0.9]
  5390. receiving
  5391. [0130 17:56:40 @multigpu.py:323] [p0574]  step: count(1110), step_time 1461.83, mean_step_time 1453.02, it/s 0.69
  5392. [0130 17:56:42 @multigpu.py:323] [p0115]  step: count(1096), step_time 1436.98, mean_step_time 1444.77, it/s 0.69
  5393. [0130 17:56:42 @multigpu.py:323] [p0576]  step: count(1120), step_time 1495.99, mean_step_time 1455.98, it/s 0.69
  5394. [0130 17:56:42 @multigpu.py:323] [p0574]  step: count(1111), step_time 1451.65, mean_step_time 1453.18, it/s 0.69
  5395. [0130 17:56:43 @multigpu.py:323] [p0115]  step: count(1097), step_time 1431.01, mean_step_time 1444.16, it/s 0.69
  5396. [0130 17:56:43 @multigpu.py:323] [p0576]  step: count(1121), step_time 1425.27, mean_step_time 1455.95, it/s 0.69
  5397. [0130 17:56:43 @multigpu.py:323] [p0574]  step: count(1112), step_time 1480.7, mean_step_time 1457.26, it/s 0.69
  5398. [0130 17:56:44 @multigpu.py:323] [p0115]  step: count(1098), step_time 1418.93, mean_step_time 1444.31, it/s 0.69
  5399. [0130 17:56:45 @multigpu.py:323] [p0576]  step: count(1122), step_time 1429.95, mean_step_time 1457.49, it/s 0.69
  5400. [0130 17:56:45 @multigpu.py:323] [p0574]  step: count(1113), step_time 1467.23, mean_step_time 1458.16, it/s 0.69
  5401. [0130 17:56:46 @multigpu.py:323] [p0115]  step: count(1099), step_time 1460.5, mean_step_time 1445.73, it/s 0.69
  5402. [0130 17:56:46 @multigpu.py:323] [p0576]  step: count(1123), step_time 1399.13, mean_step_time 1454.38, it/s 0.69
  5403. [0130 17:56:46 @multigpu.py:323] [p0574]  step: count(1114), step_time 1568.11, mean_step_time 1466.76, it/s 0.68
  5404. [0130 17:56:47 @multigpu.py:323] [p0115]  step: count(1100), step_time 1442.36, mean_step_time 1446.2, it/s 0.69
  5405. sending debugging info...
  5406. sending to address tcp://p0112:61216
  5407. ##### Sending to neptune:  mean_delay :  0.446297637489 , 0.0 #####
  5408. sending to address tcp://p0112:61216
  5409. ##### Sending to neptune:  max_delay :  0.446297637489 , -0.0 #####
  5410. ##### Sending to neptune:  min_delay :  0.446297637489 , -0.0 #####
  5411. [u'delays', [0.0, -0.0, -0.0]]
  5412. receiving
  5413. ##### Sending to neptune:  cost :  0.446298163599 , -0.00521783763543 #####
  5414. sending to address tcp://p0112:61216
  5415. ##### Sending to neptune:  policy_loss :  0.446298163599 , 0.202429369092 #####
  5416. ##### Sending to neptune:  xentropy_loss :  0.446298163599 , -2.28854560852 #####
  5417. ##### Sending to neptune:  value_loss :  0.446298163599 , 1.4182331562 #####
  5418. ##### Sending to neptune:  advantage :  0.446298163599 , -0.000883870001417 #####
  5419. ##### Sending to neptune:  pred_reward :  0.446298163599 , 0.415664792061 #####
  5420. ##### Sending to neptune:  max_logit :  0.446298163599 , 0.193621426821 #####
  5421. [u'loss', -0.005217837635427713, 0.2024293690919876, -2.288545608520508, 1.4182331562042236, -0.0008838700014166534, 0.41566479206085205, 0.193621426820755]
  5422. receiving
  5423. ##### Sending to neptune:  active_relus :  0.446298656927 , 9437430.47 #####
  5424. ##### Sending to neptune:  dp_per_s :  0.446298656927 , 89.0110297975 #####
  5425. [u'other', 9437430.47, 89.01102979748308]
  5426. receiving
  5427. [0130 17:56:47 @multigpu.py:323] [p0576]  step: count(1124), step_time 1413.07, mean_step_time 1454.27, it/s 0.69
  5428. [0130 17:56:48 @multigpu.py:323] [p0574]  step: count(1115), step_time 1509.03, mean_step_time 1472.34, it/s 0.68
  5429. [0130 17:56:49 @multigpu.py:323] [p0115]  step: count(1101), step_time 1461.1, mean_step_time 1447.2, it/s 0.69
  5430. [0130 17:56:49 @multigpu.py:323] [p0576]  step: count(1125), step_time 1438.87, mean_step_time 1455.32, it/s 0.69
  5431. [0130 17:56:49 @multigpu.py:323] [p0574]  step: count(1116), step_time 1484.95, mean_step_time 1473.62, it/s 0.68
  5432. [0130 17:56:50 @multigpu.py:323] [p0576]  step: count(1126), step_time 1414.05, mean_step_time 1454.01, it/s 0.69
  5433. [0130 17:56:50 @multigpu.py:323] [p0115]  step: count(1102), step_time 1447.27, mean_step_time 1448.57, it/s 0.69
  5434. [0130 17:56:51 @multigpu.py:323] [p0574]  step: count(1117), step_time 1576.75, mean_step_time 1480.0, it/s 0.68
  5435. [0130 17:56:52 @multigpu.py:323] [p0115]  step: count(1103), step_time 1403.75, mean_step_time 1447.6, it/s 0.69
  5436. [0130 17:56:52 @multigpu.py:323] [p0576]  step: count(1127), step_time 1415.42, mean_step_time 1451.52, it/s 0.69
  5437. [0130 17:56:52 @multigpu.py:323] [p0574]  step: count(1118), step_time 1448.47, mean_step_time 1476.25, it/s 0.68
  5438. [0130 17:56:53 @multigpu.py:323] [p0115]  step: count(1104), step_time 1416.13, mean_step_time 1444.54, it/s 0.69
  5439. [0130 17:56:53 @multigpu.py:323] [p0576]  step: count(1128), step_time 1455.43, mean_step_time 1454.13, it/s 0.69
  5440. [0130 17:56:54 @multigpu.py:323] [p0574]  step: count(1119), step_time 1405.29, mean_step_time 1472.9, it/s 0.68
  5441. [0130 17:56:54 @multigpu.py:323] [p0115]  step: count(1105), step_time 1434.33, mean_step_time 1446.67, it/s 0.69
  5442. [0130 17:56:54 @multigpu.py:323] [p0576]  step: count(1129), step_time 1427.13, mean_step_time 1454.73, it/s 0.69
  5443. [0130 17:56:55 @multigpu.py:323] [p0574]  step: count(1120), step_time 1456.45, mean_step_time 1475.27, it/s 0.68
  5444. [0130 17:56:56 @multigpu.py:323] [p0115]  step: count(1106), step_time 1438.43, mean_step_time 1446.99, it/s 0.69
  5445. [0130 17:56:56 @multigpu.py:323] [p0576]  step: count(1130), step_time 1414.84, mean_step_time 1453.57, it/s 0.69
  5446. [0130 17:56:57 @multigpu.py:323] [p0574]  step: count(1121), step_time 1412.17, mean_step_time 1473.72, it/s 0.68
  5447. [0130 17:56:57 @multigpu.py:323] [p0115]  step: count(1107), step_time 1409.18, mean_step_time 1445.67, it/s 0.69
  5448. [0130 17:56:57 @multigpu.py:323] [p0576]  step: count(1131), step_time 1397.04, mean_step_time 1451.85, it/s 0.69
  5449. [0130 17:56:58 @multigpu.py:323] [p0574]  step: count(1122), step_time 1509.22, mean_step_time 1477.5, it/s 0.68
  5450. [0130 17:57:00 @multigpu.py:323] [p0574]  step: count(1123), step_time 1393.09, mean_step_time 1471.51, it/s 0.68
  5451. [0130 17:57:00 @multigpu.py:323] [p0115]  step: count(1108), step_time 2220.94, mean_step_time 1471.64, it/s 0.68
  5452. [0130 17:57:00 @multigpu.py:323] [p0576]  step: count(1132), step_time 2216.0, mean_step_time 1471.52, it/s 0.68
  5453. [0130 17:57:01 @multigpu.py:323] [p0115]  step: count(1109), step_time 1432.01, mean_step_time 1470.62, it/s 0.68
  5454. [0130 17:57:01 @multigpu.py:323] [p0576]  step: count(1133), step_time 1447.93, mean_step_time 1470.58, it/s 0.68
  5455. [0130 17:57:01 @multigpu.py:323] [p0574]  step: count(1124), step_time 1464.41, mean_step_time 1468.09, it/s 0.68
  5456. [0130 17:57:02 @multigpu.py:323] [p0574]  step: count(1125), step_time 1372.86, mean_step_time 1463.66, it/s 0.68
  5457. [0130 17:57:02 @multigpu.py:323] [p0115]  step: count(1110), step_time 1438.77, mean_step_time 1473.34, it/s 0.68
  5458. [0130 17:57:02 @multigpu.py:323] [p0576]  step: count(1134), step_time 1478.64, mean_step_time 1472.95, it/s 0.68
  5459. [0130 17:57:04 @multigpu.py:323] [p0115]  step: count(1111), step_time 1376.28, mean_step_time 1470.66, it/s 0.68
  5460. [0130 17:57:04 @multigpu.py:323] [p0574]  step: count(1126), step_time 1469.52, mean_step_time 1462.75, it/s 0.68
  5461. [0130 17:57:04 @multigpu.py:323] [p0576]  step: count(1135), step_time 1400.41, mean_step_time 1472.81, it/s 0.68
  5462. sending to address tcp://p0112:61216
  5463. ##### Sending to neptune:  online_score :  0.451079662442 , 2.5 #####
  5464. [u'online', 2.5]
  5465. ##### Sending to neptune:  active_workers :  0.451079749399 , 3 #####
  5466. receiving
  5467. [0130 17:57:05 @multigpu.py:323] [p0574]  step: count(1127), step_time 1377.8, mean_step_time 1462.32, it/s 0.68
  5468. [0130 17:57:05 @multigpu.py:323] [p0115]  step: count(1112), step_time 1473.36, mean_step_time 1471.92, it/s 0.68
  5469. [0130 17:57:05 @multigpu.py:323] [p0576]  step: count(1136), step_time 1422.99, mean_step_time 1472.11, it/s 0.68
  5470. sending to address tcp://p0112:61216
  5471. ##### Sending to neptune:  online_score :  0.451481641663 , 1.4 #####
  5472. [u'online', 1.4]
  5473. receiving
  5474. [0130 17:57:07 @multigpu.py:323] [p0574]  step: count(1128), step_time 1433.09, mean_step_time 1458.19, it/s 0.69
  5475. [0130 17:57:07 @multigpu.py:323] [p0115]  step: count(1113), step_time 1415.25, mean_step_time 1472.6, it/s 0.68
  5476. [0130 17:57:07 @multigpu.py:323] [p0576]  step: count(1137), step_time 1427.53, mean_step_time 1472.22, it/s 0.68
  5477. [0130 17:57:08 @multigpu.py:323] [p0574]  step: count(1129), step_time 1434.12, mean_step_time 1458.84, it/s 0.69
  5478. [0130 17:57:08 @multigpu.py:323] [p0576]  step: count(1138), step_time 1416.3, mean_step_time 1470.62, it/s 0.68
  5479. [0130 17:57:08 @multigpu.py:323] [p0115]  step: count(1114), step_time 1460.82, mean_step_time 1472.78, it/s 0.68
  5480. [0130 17:57:09 @multigpu.py:323] [p0574]  step: count(1130), step_time 1433.63, mean_step_time 1457.43, it/s 0.69
  5481. [0130 17:57:10 @multigpu.py:323] [p0576]  step: count(1139), step_time 1407.76, mean_step_time 1467.19, it/s 0.68
  5482. [0130 17:57:10 @multigpu.py:323] [p0115]  step: count(1115), step_time 1432.65, mean_step_time 1472.5, it/s 0.68
  5483. [0130 17:57:11 @multigpu.py:323] [p0574]  step: count(1131), step_time 1405.0, mean_step_time 1455.09, it/s 0.69
  5484. [0130 17:57:11 @multigpu.py:323] [p0576]  step: count(1140), step_time 1421.67, mean_step_time 1463.47, it/s 0.68
  5485. [0130 17:57:11 @multigpu.py:323] [p0115]  step: count(1116), step_time 1431.55, mean_step_time 1472.23, it/s 0.68
  5486. [0130 17:57:12 @multigpu.py:323] [p0574]  step: count(1132), step_time 1475.83, mean_step_time 1454.85, it/s 0.69
  5487. [0130 17:57:12 @multigpu.py:323] [p0576]  step: count(1141), step_time 1445.96, mean_step_time 1464.51, it/s 0.68
  5488. [0130 17:57:12 @multigpu.py:323] [p0115]  step: count(1117), step_time 1491.65, mean_step_time 1475.26, it/s 0.68
  5489. [0130 17:57:14 @multigpu.py:323] [p0574]  step: count(1133), step_time 1395.66, mean_step_time 1451.27, it/s 0.69
  5490. [0130 17:57:14 @multigpu.py:323] [p0576]  step: count(1142), step_time 1430.69, mean_step_time 1464.54, it/s 0.68
  5491. [0130 17:57:14 @multigpu.py:323] [p0115]  step: count(1118), step_time 1422.91, mean_step_time 1475.46, it/s 0.68
  5492. [0130 17:57:15 @multigpu.py:323] [p0576]  step: count(1143), step_time 1401.77, mean_step_time 1464.68, it/s 0.68
  5493. [0130 17:57:15 @multigpu.py:323] [p0574]  step: count(1134), step_time 1496.56, mean_step_time 1447.69, it/s 0.69
  5494. [0130 17:57:15 @multigpu.py:323] [p0115]  step: count(1119), step_time 1423.35, mean_step_time 1473.6, it/s 0.68
  5495. [0130 17:57:17 @multigpu.py:323] [p0576]  step: count(1144), step_time 1392.55, mean_step_time 1463.65, it/s 0.68
  5496. [0130 17:57:17 @multigpu.py:323] [p0574]  step: count(1135), step_time 1422.67, mean_step_time 1443.38, it/s 0.69
  5497. [0130 17:57:17 @multigpu.py:323] [p0115]  step: count(1120), step_time 1431.39, mean_step_time 1473.06, it/s 0.68
  5498. sending to address tcp://p0112:61216
  5499. ##### Sending to neptune:  online_score :  0.454599559704 , 2.2 #####
  5500. [u'online', 2.2]
  5501. receiving
  5502. [0130 17:57:18 @multigpu.py:323] [p0576]  step: count(1145), step_time 1391.93, mean_step_time 1461.3, it/s 0.68
  5503. [0130 17:57:18 @multigpu.py:323] [p0115]  step: count(1121), step_time 1430.98, mean_step_time 1471.55, it/s 0.68
  5504. [0130 17:57:18 @multigpu.py:323] [p0574]  step: count(1136), step_time 1487.91, mean_step_time 1443.52, it/s 0.69
  5505. [0130 17:57:20 @multigpu.py:323] [p0576]  step: count(1146), step_time 1514.51, mean_step_time 1466.33, it/s 0.68
  5506. [0130 17:57:20 @multigpu.py:323] [p0574]  step: count(1137), step_time 1420.15, mean_step_time 1435.69, it/s 0.7
  5507. [0130 17:57:20 @multigpu.py:323] [p0115]  step: count(1122), step_time 1458.52, mean_step_time 1472.11, it/s 0.68
  5508. [0130 17:57:21 @multigpu.py:323] [p0576]  step: count(1147), step_time 1380.34, mean_step_time 1464.57, it/s 0.68
  5509. [0130 17:57:21 @multigpu.py:323] [p0574]  step: count(1138), step_time 1355.25, mean_step_time 1431.03, it/s 0.7
  5510. [0130 17:57:21 @multigpu.py:323] [p0115]  step: count(1123), step_time 1428.35, mean_step_time 1473.34, it/s 0.68
  5511. [0130 17:57:22 @multigpu.py:323] [p0576]  step: count(1148), step_time 1437.57, mean_step_time 1463.68, it/s 0.68
  5512. [0130 17:57:22 @multigpu.py:323] [p0574]  step: count(1139), step_time 1437.97, mean_step_time 1432.67, it/s 0.7
  5513. [0130 17:57:22 @multigpu.py:323] [p0115]  step: count(1124), step_time 1431.56, mean_step_time 1474.11, it/s 0.68
  5514. [0130 17:57:24 @multigpu.py:323] [p0576]  step: count(1149), step_time 1432.13, mean_step_time 1463.93, it/s 0.68
  5515. [0130 17:57:24 @multigpu.py:323] [p0574]  step: count(1140), step_time 1433.31, mean_step_time 1431.51, it/s 0.7
  5516. [0130 17:57:24 @multigpu.py:323] [p0115]  step: count(1125), step_time 1474.25, mean_step_time 1476.11, it/s 0.68
  5517. [0130 17:57:25 @multigpu.py:323] [p0576]  step: count(1150), step_time 1435.1, mean_step_time 1464.94, it/s 0.68
  5518. [0130 17:57:25 @multigpu.py:323] [p0574]  step: count(1141), step_time 1379.84, mean_step_time 1429.89, it/s 0.7
  5519. [0130 17:57:25 @multigpu.py:323] [p0115]  step: count(1126), step_time 1453.67, mean_step_time 1476.87, it/s 0.68
  5520. [0130 17:57:27 @multigpu.py:323] [p0576]  step: count(1151), step_time 1389.31, mean_step_time 1464.55, it/s 0.68
  5521. [0130 17:57:27 @multigpu.py:323] [p0574]  step: count(1142), step_time 1416.3, mean_step_time 1425.25, it/s 0.7
  5522. [0130 17:57:27 @multigpu.py:323] [p0115]  step: count(1127), step_time 1405.31, mean_step_time 1476.68, it/s 0.68
  5523. [0130 17:57:28 @multigpu.py:323] [p0115]  step: count(1128), step_time 1483.4, mean_step_time 1439.8, it/s 0.69
  5524. [0130 17:57:28 @multigpu.py:323] [p0574]  step: count(1143), step_time 1683.21, mean_step_time 1439.75, it/s 0.69
  5525. [0130 17:57:28 @multigpu.py:323] [p0576]  step: count(1152), step_time 1719.88, mean_step_time 1439.75, it/s 0.69
  5526. [0130 17:57:30 @multigpu.py:323] [p0115]  step: count(1129), step_time 1378.47, mean_step_time 1437.12, it/s 0.7
  5527. [0130 17:57:30 @multigpu.py:323] [p0576]  step: count(1153), step_time 1401.52, mean_step_time 1437.43, it/s 0.7
  5528. [0130 17:57:30 @multigpu.py:323] [p0574]  step: count(1144), step_time 1404.0, mean_step_time 1436.73, it/s 0.7
  5529. [0130 17:57:31 @multigpu.py:323] [p0115]  step: count(1130), step_time 1440.34, mean_step_time 1437.2, it/s 0.7
  5530. [0130 17:57:31 @multigpu.py:323] [p0574]  step: count(1145), step_time 1415.34, mean_step_time 1438.86, it/s 0.69
  5531. [0130 17:57:31 @multigpu.py:323] [p0576]  step: count(1154), step_time 1429.49, mean_step_time 1434.97, it/s 0.7
  5532. [0130 17:57:33 @multigpu.py:323] [p0576]  step: count(1155), step_time 1406.06, mean_step_time 1435.25, it/s 0.7
  5533. [0130 17:57:33 @multigpu.py:323] [p0574]  step: count(1146), step_time 1434.59, mean_step_time 1437.11, it/s 0.7
  5534. [0130 17:57:33 @multigpu.py:323] [p0115]  step: count(1131), step_time 1466.16, mean_step_time 1441.7, it/s 0.69
  5535. sending to address tcp://p0112:61216
  5536. ##### Sending to neptune:  online_score :  0.459099557201 , 2.4 #####
  5537. [u'online', 2.4]
  5538. receiving
  5539. [0130 17:57:34 @multigpu.py:323] [p0115]  step: count(1132), step_time 1373.58, mean_step_time 1436.71, it/s 0.7
  5540. [0130 17:57:34 @multigpu.py:323] [p0576]  step: count(1156), step_time 1426.12, mean_step_time 1435.41, it/s 0.7
  5541. [0130 17:57:34 @multigpu.py:323] [p0574]  step: count(1147), step_time 1435.5, mean_step_time 1440.0, it/s 0.69
  5542. sending to address tcp://p0112:61216
  5543. ##### Sending to neptune:  online_score :  0.459456433588 , 1.6 #####
  5544. [u'online', 1.6]
  5545. receiving
  5546. [0130 17:57:35 @multigpu.py:323] [p0574]  step: count(1148), step_time 1371.35, mean_step_time 1436.91, it/s 0.7
  5547. [0130 17:57:35 @multigpu.py:323] [p0576]  step: count(1157), step_time 1404.18, mean_step_time 1434.24, it/s 0.7
  5548. [0130 17:57:35 @multigpu.py:323] [p0115]  step: count(1133), step_time 1432.68, mean_step_time 1437.58, it/s 0.7
  5549. [0130 17:57:37 @multigpu.py:323] [p0574]  step: count(1149), step_time 1419.67, mean_step_time 1436.19, it/s 0.7
  5550. [0130 17:57:37 @multigpu.py:323] [p0115]  step: count(1134), step_time 1433.33, mean_step_time 1436.21, it/s 0.7
  5551. [0130 17:57:37 @multigpu.py:323] [p0576]  step: count(1158), step_time 1456.34, mean_step_time 1436.24, it/s 0.7
  5552. [0130 17:57:38 @multigpu.py:323] [p0574]  step: count(1150), step_time 1417.75, mean_step_time 1435.39, it/s 0.7
  5553. [0130 17:57:38 @multigpu.py:323] [p0115]  step: count(1135), step_time 1424.22, mean_step_time 1435.78, it/s 0.7
  5554. [0130 17:57:38 @multigpu.py:323] [p0576]  step: count(1159), step_time 1440.73, mean_step_time 1437.89, it/s 0.7
  5555. [0130 17:57:40 @multigpu.py:323] [p0574]  step: count(1151), step_time 1422.46, mean_step_time 1436.27, it/s 0.7
  5556. [0130 17:57:40 @multigpu.py:323] [p0115]  step: count(1136), step_time 1428.24, mean_step_time 1435.62, it/s 0.7
  5557. [0130 17:57:40 @multigpu.py:323] [p0576]  step: count(1160), step_time 1450.56, mean_step_time 1439.34, it/s 0.69
  5558. [0130 17:57:41 @multigpu.py:323] [p0576]  step: count(1161), step_time 1392.22, mean_step_time 1436.65, it/s 0.7
  5559. [0130 17:57:41 @multigpu.py:323] [p0574]  step: count(1152), step_time 1491.29, mean_step_time 1437.04, it/s 0.7
  5560. [0130 17:57:41 @multigpu.py:323] [p0115]  step: count(1137), step_time 1469.56, mean_step_time 1434.51, it/s 0.7
  5561. [0130 17:57:43 @multigpu.py:323] [p0574]  step: count(1153), step_time 1377.57, mean_step_time 1436.13, it/s 0.7
  5562. [0130 17:57:43 @multigpu.py:323] [p0576]  step: count(1162), step_time 1424.6, mean_step_time 1436.35, it/s 0.7
  5563. [0130 17:57:43 @multigpu.py:323] [p0115]  step: count(1138), step_time 1468.64, mean_step_time 1436.8, it/s 0.7
  5564. [0130 17:57:44 @multigpu.py:323] [p0574]  step: count(1154), step_time 1434.21, mean_step_time 1433.02, it/s 0.7
  5565. [0130 17:57:44 @multigpu.py:323] [p0576]  step: count(1163), step_time 1407.8, mean_step_time 1436.65, it/s 0.7
  5566. [0130 17:57:44 @multigpu.py:323] [p0115]  step: count(1139), step_time 1388.54, mean_step_time 1435.06, it/s 0.7
  5567. [0130 17:57:45 @multigpu.py:323] [p0574]  step: count(1155), step_time 1405.23, mean_step_time 1432.14, it/s 0.7
  5568. [0130 17:57:45 @multigpu.py:323] [p0576]  step: count(1164), step_time 1430.82, mean_step_time 1438.56, it/s 0.7
  5569. [0130 17:57:46 @multigpu.py:323] [p0115]  step: count(1140), step_time 1482.05, mean_step_time 1437.59, it/s 0.7
  5570. [0130 17:57:47 @multigpu.py:323] [p0574]  step: count(1156), step_time 1441.58, mean_step_time 1429.83, it/s 0.7
  5571. [0130 17:57:47 @multigpu.py:323] [p0576]  step: count(1165), step_time 1415.85, mean_step_time 1439.76, it/s 0.69
  5572. [0130 17:57:47 @multigpu.py:323] [p0115]  step: count(1141), step_time 1376.95, mean_step_time 1434.89, it/s 0.7
  5573. sending to address tcp://p0112:61216
  5574. ##### Sending to neptune:  online_score :  0.462921938035 , 1.9 #####
  5575. [u'online', 1.9]
  5576. receiving
  5577. [0130 17:57:48 @multigpu.py:323] [p0576]  step: count(1166), step_time 1386.7, mean_step_time 1433.36, it/s 0.7
  5578. [0130 17:57:48 @multigpu.py:323] [p0574]  step: count(1157), step_time 1514.69, mean_step_time 1434.56, it/s 0.7
  5579. [0130 17:57:48 @multigpu.py:323] [p0115]  step: count(1142), step_time 1424.27, mean_step_time 1433.18, it/s 0.7
  5580. [0130 17:57:50 @multigpu.py:323] [p0576]  step: count(1167), step_time 1497.67, mean_step_time 1439.23, it/s 0.69
  5581. [0130 17:57:50 @multigpu.py:323] [p0574]  step: count(1158), step_time 1395.61, mean_step_time 1436.57, it/s 0.7
  5582. [0130 17:57:50 @multigpu.py:323] [p0115]  step: count(1143), step_time 1457.88, mean_step_time 1434.66, it/s 0.7
  5583. [0130 17:57:51 @multigpu.py:323] [p0576]  step: count(1168), step_time 1398.13, mean_step_time 1437.26, it/s 0.7
  5584. [0130 17:57:51 @multigpu.py:323] [p0574]  step: count(1159), step_time 1416.04, mean_step_time 1435.48, it/s 0.7
  5585. [0130 17:57:51 @multigpu.py:323] [p0115]  step: count(1144), step_time 1414.34, mean_step_time 1433.79, it/s 0.7
  5586. [0130 17:57:53 @multigpu.py:323] [p0576]  step: count(1169), step_time 1466.74, mean_step_time 1438.99, it/s 0.69
  5587. [0130 17:57:53 @multigpu.py:323] [p0574]  step: count(1160), step_time 1441.19, mean_step_time 1435.87, it/s 0.7
  5588. [0130 17:57:53 @multigpu.py:323] [p0115]  step: count(1145), step_time 1470.76, mean_step_time 1433.62, it/s 0.7
  5589. [0130 17:57:54 @multigpu.py:323] [p0574]  step: count(1161), step_time 1391.38, mean_step_time 1436.45, it/s 0.7
  5590. [0130 17:57:54 @multigpu.py:323] [p0576]  step: count(1170), step_time 1394.21, mean_step_time 1436.95, it/s 0.7
  5591. [0130 17:57:54 @multigpu.py:323] [p0115]  step: count(1146), step_time 1397.14, mean_step_time 1430.79, it/s 0.7
  5592. [0130 17:57:55 @multigpu.py:323] [p0574]  step: count(1162), step_time 1442.57, mean_step_time 1437.76, it/s 0.7
  5593. [0130 17:57:55 @multigpu.py:323] [p0576]  step: count(1171), step_time 1501.38, mean_step_time 1442.55, it/s 0.69
  5594. [0130 17:57:55 @multigpu.py:323] [p0115]  step: count(1147), step_time 1421.85, mean_step_time 1431.62, it/s 0.7
  5595. [0130 17:57:57 @multigpu.py:323] [p0115]  step: count(1148), step_time 1466.54, mean_step_time 1430.78, it/s 0.7
  5596. [0130 17:57:57 @multigpu.py:323] [p0576]  step: count(1172), step_time 1483.5, mean_step_time 1430.73, it/s 0.7
  5597. [0130 17:57:57 @multigpu.py:323] [p0574]  step: count(1163), step_time 1542.55, mean_step_time 1430.73, it/s 0.7
  5598. [0130 17:57:58 @multigpu.py:323] [p0115]  step: count(1149), step_time 1364.25, mean_step_time 1430.07, it/s 0.7
  5599. [0130 17:57:58 @multigpu.py:323] [p0574]  step: count(1164), step_time 1429.55, mean_step_time 1432.01, it/s 0.7
  5600. [0130 17:57:58 @multigpu.py:323] [p0576]  step: count(1173), step_time 1446.76, mean_step_time 1432.99, it/s 0.7
  5601. sending to address tcp://p0112:61216
  5602. ##### Sending to neptune:  online_score :  0.466300260226 , 1.6 #####
  5603. [u'online', 1.6]
  5604. receiving
  5605. [0130 17:58:00 @multigpu.py:323] [p0115]  step: count(1150), step_time 1485.05, mean_step_time 1432.3, it/s 0.7
  5606. [0130 17:58:00 @multigpu.py:323] [p0576]  step: count(1174), step_time 1423.04, mean_step_time 1432.67, it/s 0.7
  5607. [0130 17:58:00 @multigpu.py:323] [p0574]  step: count(1165), step_time 1455.67, mean_step_time 1434.02, it/s 0.7
  5608. [0130 17:58:01 @multigpu.py:323] [p0574]  step: count(1166), step_time 1392.81, mean_step_time 1431.93, it/s 0.7
  5609. [0130 17:58:01 @multigpu.py:323] [p0115]  step: count(1151), step_time 1451.4, mean_step_time 1431.56, it/s 0.7
  5610. [0130 17:58:01 @multigpu.py:323] [p0576]  step: count(1175), step_time 1432.38, mean_step_time 1433.99, it/s 0.7
  5611. [0130 17:58:03 @multigpu.py:323] [p0574]  step: count(1167), step_time 1418.52, mean_step_time 1431.09, it/s 0.7
  5612. [0130 17:58:03 @multigpu.py:323] [p0115]  step: count(1152), step_time 1432.02, mean_step_time 1434.49, it/s 0.7
  5613. [0130 17:58:03 @multigpu.py:323] [p0576]  step: count(1176), step_time 1450.0, mean_step_time 1435.18, it/s 0.7
  5614. [0130 17:58:04 @multigpu.py:323] [p0574]  step: count(1168), step_time 1430.0, mean_step_time 1434.02, it/s 0.7
  5615. [0130 17:58:04 @multigpu.py:323] [p0115]  step: count(1153), step_time 1400.37, mean_step_time 1432.87, it/s 0.7
  5616. [0130 17:58:04 @multigpu.py:323] [p0576]  step: count(1177), step_time 1410.86, mean_step_time 1435.51, it/s 0.7
  5617. [0130 17:58:06 @multigpu.py:323] [p0574]  step: count(1169), step_time 1430.97, mean_step_time 1434.58, it/s 0.7
  5618. [0130 17:58:06 @multigpu.py:323] [p0576]  step: count(1178), step_time 1409.64, mean_step_time 1433.18, it/s 0.7
  5619. [0130 17:58:06 @multigpu.py:323] [p0115]  step: count(1154), step_time 1457.51, mean_step_time 1434.08, it/s 0.7
  5620. [0130 17:58:07 @multigpu.py:323] [p0574]  step: count(1170), step_time 1406.46, mean_step_time 1434.02, it/s 0.7
  5621. [0130 17:58:07 @multigpu.py:323] [p0576]  step: count(1179), step_time 1408.16, mean_step_time 1431.55, it/s 0.7
  5622. [0130 17:58:07 @multigpu.py:323] [p0115]  step: count(1155), step_time 1438.62, mean_step_time 1434.8, it/s 0.7
  5623. [0130 17:58:08 @multigpu.py:323] [p0576]  step: count(1180), step_time 1363.05, mean_step_time 1427.18, it/s 0.7
  5624. [0130 17:58:08 @multigpu.py:323] [p0574]  step: count(1171), step_time 1409.52, mean_step_time 1433.37, it/s 0.7
  5625. [0130 17:58:08 @multigpu.py:323] [p0115]  step: count(1156), step_time 1448.35, mean_step_time 1435.81, it/s 0.7
  5626. sending to address tcp://p0112:61216
  5627. ##### Sending to neptune:  online_score :  0.468956205514 , 1.7 #####
  5628. [u'online', 1.7]
  5629. ##### Sending to neptune:  active_workers :  0.468956336644 , 3 #####
  5630. receiving
  5631. [0130 17:58:10 @multigpu.py:323] [p0574]  step: count(1172), step_time 1419.6, mean_step_time 1429.79, it/s 0.7
  5632. [0130 17:58:10 @multigpu.py:323] [p0576]  step: count(1181), step_time 1521.66, mean_step_time 1433.65, it/s 0.7
  5633. [0130 17:58:10 @multigpu.py:323] [p0115]  step: count(1157), step_time 1429.55, mean_step_time 1433.81, it/s 0.7
  5634. [0130 17:58:11 @multigpu.py:323] [p0574]  step: count(1173), step_time 1444.48, mean_step_time 1433.13, it/s 0.7
  5635. [0130 17:58:11 @multigpu.py:323] [p0115]  step: count(1158), step_time 1404.46, mean_step_time 1430.6, it/s 0.7
  5636. [0130 17:58:11 @multigpu.py:323] [p0576]  step: count(1182), step_time 1444.9, mean_step_time 1434.66, it/s 0.7
  5637. [0130 17:58:13 @multigpu.py:323] [p0574]  step: count(1174), step_time 1458.27, mean_step_time 1434.33, it/s 0.7
  5638. [0130 17:58:13 @multigpu.py:323] [p0576]  step: count(1183), step_time 1409.56, mean_step_time 1434.75, it/s 0.7
  5639. [0130 17:58:13 @multigpu.py:323] [p0115]  step: count(1159), step_time 1438.83, mean_step_time 1433.11, it/s 0.7
  5640. [0130 17:58:14 @multigpu.py:323] [p0574]  step: count(1175), step_time 1425.11, mean_step_time 1435.33, it/s 0.7
  5641. [0130 17:58:14 @multigpu.py:323] [p0576]  step: count(1184), step_time 1405.42, mean_step_time 1433.48, it/s 0.7
  5642. [0130 17:58:14 @multigpu.py:323] [p0115]  step: count(1160), step_time 1442.89, mean_step_time 1431.15, it/s 0.7
  5643. [0130 17:58:15 @multigpu.py:323] [p0576]  step: count(1185), step_time 1369.21, mean_step_time 1431.15, it/s 0.7
  5644. [0130 17:58:16 @multigpu.py:323] [p0574]  step: count(1176), step_time 1445.12, mean_step_time 1435.51, it/s 0.7
  5645. [0130 17:58:16 @multigpu.py:323] [p0115]  step: count(1161), step_time 1429.14, mean_step_time 1433.76, it/s 0.7
  5646. [0130 17:58:17 @multigpu.py:323] [p0576]  step: count(1186), step_time 1430.49, mean_step_time 1433.34, it/s 0.7
  5647. [0130 17:58:17 @multigpu.py:323] [p0574]  step: count(1177), step_time 1413.46, mean_step_time 1430.44, it/s 0.7
  5648. [0130 17:58:17 @multigpu.py:323] [p0115]  step: count(1162), step_time 1433.96, mean_step_time 1434.25, it/s 0.7
  5649. [0130 17:58:18 @multigpu.py:323] [p0576]  step: count(1187), step_time 1365.08, mean_step_time 1426.71, it/s 0.7
  5650. [0130 17:58:18 @multigpu.py:323] [p0574]  step: count(1178), step_time 1520.99, mean_step_time 1436.71, it/s 0.7
  5651. [0130 17:58:18 @multigpu.py:323] [p0115]  step: count(1163), step_time 1451.15, mean_step_time 1433.91, it/s 0.7
  5652. [0130 17:58:20 @multigpu.py:323] [p0576]  step: count(1188), step_time 1456.22, mean_step_time 1429.61, it/s 0.7
  5653. [0130 17:58:20 @multigpu.py:323] [p0574]  step: count(1179), step_time 1405.58, mean_step_time 1436.19, it/s 0.7
  5654. [0130 17:58:20 @multigpu.py:323] [p0115]  step: count(1164), step_time 1411.96, mean_step_time 1433.79, it/s 0.7
  5655. [0130 17:58:21 @multigpu.py:323] [p0576]  step: count(1189), step_time 1377.94, mean_step_time 1425.17, it/s 0.7
  5656. [0130 17:58:21 @multigpu.py:323] [p0115]  step: count(1165), step_time 1390.93, mean_step_time 1429.8, it/s 0.7
  5657. [0130 17:58:21 @multigpu.py:323] [p0574]  step: count(1180), step_time 1425.96, mean_step_time 1435.43, it/s 0.7
  5658. [0130 17:58:22 @multigpu.py:323] [p0576]  step: count(1190), step_time 1423.47, mean_step_time 1426.64, it/s 0.7
  5659. [0130 17:58:23 @multigpu.py:323] [p0115]  step: count(1166), step_time 1432.86, mean_step_time 1431.59, it/s 0.7
  5660. [0130 17:58:23 @multigpu.py:323] [p0574]  step: count(1181), step_time 1417.27, mean_step_time 1436.72, it/s 0.7
  5661. [0130 17:58:24 @multigpu.py:323] [p0576]  step: count(1191), step_time 1413.47, mean_step_time 1422.24, it/s 0.7
  5662. sending to address tcp://p0112:61216
  5663. ##### Sending to neptune:  online_score :  0.473143153853 , 2.2 #####
  5664. [u'online', 2.2]
  5665. receiving
  5666. [0130 17:58:24 @multigpu.py:323] [p0115]  step: count(1167), step_time 1428.48, mean_step_time 1431.92, it/s 0.7
  5667. [0130 17:58:24 @multigpu.py:323] [p0574]  step: count(1182), step_time 1495.99, mean_step_time 1439.39, it/s 0.69
  5668. [0130 17:58:26 @multigpu.py:323] [p0115]  step: count(1168), step_time 1556.94, mean_step_time 1436.44, it/s 0.7
  5669. [0130 17:58:26 @multigpu.py:323] [p0574]  step: count(1183), step_time 1481.8, mean_step_time 1436.36, it/s 0.7
  5670. [0130 17:58:26 @multigpu.py:323] [p0576]  step: count(1192), step_time 1767.82, mean_step_time 1436.46, it/s 0.7
  5671. [0130 17:58:27 @multigpu.py:323] [p0574]  step: count(1184), step_time 1398.17, mean_step_time 1434.79, it/s 0.7
  5672. [0130 17:58:27 @multigpu.py:323] [p0115]  step: count(1169), step_time 1428.32, mean_step_time 1439.64, it/s 0.69
  5673. [0130 17:58:27 @multigpu.py:323] [p0576]  step: count(1193), step_time 1467.1, mean_step_time 1437.47, it/s 0.7
  5674. [0130 17:58:29 @multigpu.py:323] [p0574]  step: count(1185), step_time 1440.13, mean_step_time 1434.01, it/s 0.7
  5675. [0130 17:58:29 @multigpu.py:323] [p0576]  step: count(1194), step_time 1377.48, mean_step_time 1435.2, it/s 0.7
  5676. [0130 17:58:29 @multigpu.py:323] [p0115]  step: count(1170), step_time 1434.16, mean_step_time 1437.1, it/s 0.7
  5677. sending to address tcp://p0112:61216
  5678. ##### Sending to neptune:  online_score :  0.474509788288 , 2.0 #####
  5679. [u'online', 2.0]
  5680. receiving
  5681. [0130 17:58:30 @multigpu.py:323] [p0574]  step: count(1186), step_time 1390.93, mean_step_time 1433.92, it/s 0.7
  5682. [0130 17:58:30 @multigpu.py:323] [p0576]  step: count(1195), step_time 1464.69, mean_step_time 1436.81, it/s 0.7
  5683. [0130 17:58:30 @multigpu.py:323] [p0115]  step: count(1171), step_time 1453.03, mean_step_time 1437.18, it/s 0.7
  5684. [0130 17:58:31 @multigpu.py:323] [p0574]  step: count(1187), step_time 1418.75, mean_step_time 1433.93, it/s 0.7
  5685. [0130 17:58:31 @multigpu.py:323] [p0115]  step: count(1172), step_time 1395.88, mean_step_time 1435.37, it/s 0.7
  5686. [0130 17:58:31 @multigpu.py:323] [p0576]  step: count(1196), step_time 1400.67, mean_step_time 1434.35, it/s 0.7
  5687. [0130 17:58:33 @multigpu.py:323] [p0574]  step: count(1188), step_time 1443.55, mean_step_time 1434.61, it/s 0.7
  5688. [0130 17:58:33 @multigpu.py:323] [p0576]  step: count(1197), step_time 1418.83, mean_step_time 1434.74, it/s 0.7
  5689. [0130 17:58:33 @multigpu.py:323] [p0115]  step: count(1173), step_time 1432.91, mean_step_time 1437.0, it/s 0.7
  5690. sending to address tcp://p0112:61216
  5691. ##### Sending to neptune:  online_score :  0.475631449156 , 1.7 #####
  5692. [u'online', 1.7]
  5693. receiving
  5694. [0130 17:58:34 @multigpu.py:323] [p0574]  step: count(1189), step_time 1408.25, mean_step_time 1433.47, it/s 0.7
  5695. [0130 17:58:34 @multigpu.py:323] [p0115]  step: count(1174), step_time 1381.5, mean_step_time 1433.2, it/s 0.7
  5696. [0130 17:58:34 @multigpu.py:323] [p0576]  step: count(1198), step_time 1453.78, mean_step_time 1436.95, it/s 0.7
  5697. [0130 17:58:36 @multigpu.py:323] [p0574]  step: count(1190), step_time 1429.62, mean_step_time 1434.63, it/s 0.7
  5698. [0130 17:58:36 @multigpu.py:323] [p0115]  step: count(1175), step_time 1427.4, mean_step_time 1432.64, it/s 0.7
  5699. [0130 17:58:36 @multigpu.py:323] [p0576]  step: count(1199), step_time 1442.31, mean_step_time 1438.66, it/s 0.7
  5700. [0130 17:58:37 @multigpu.py:323] [p0115]  step: count(1176), step_time 1407.48, mean_step_time 1430.59, it/s 0.7
  5701. [0130 17:58:37 @multigpu.py:323] [p0576]  step: count(1200), step_time 1383.31, mean_step_time 1439.67, it/s 0.69
  5702. sending debugging info...
  5703. sending to address tcp://p0112:61216
  5704. ##### Sending to neptune:  mean_delay :  0.476802181337 , 0.0 #####
  5705. sending to address tcp://p0112:61216
  5706. ##### Sending to neptune:  max_delay :  0.476802181337 , -0.0 #####
  5707. ##### Sending to neptune:  min_delay :  0.476802181337 , -0.0 #####
  5708. [u'delays', [0.0, -0.0, -0.0]]
  5709. receiving
  5710. ##### Sending to neptune:  cost :  0.476802654399 , -0.00253472337499 #####
  5711. ##### Sending to neptune:  policy_loss :  0.476802654399 , 0.375536680222 #####
  5712. sending to address tcp://p0112:61216
  5713. ##### Sending to neptune:  xentropy_loss :  0.476802654399 , -2.28917598724 #####
  5714. ##### Sending to neptune:  value_loss :  0.476802654399 , 1.58919465542 #####
  5715. ##### Sending to neptune:  advantage :  0.476802654399 , -0.00181525712833 #####
  5716. ##### Sending to neptune:  pred_reward :  0.476802654399 , 0.453048050404 #####
  5717. ##### Sending to neptune:  max_logit :  0.476802654399 , 0.191985800862 #####
  5718. [u'loss', -0.002534723374992609, 0.3755366802215576, -2.2891759872436523, 1.589194655418396, -0.001815257128328085, 0.45304805040359497, 0.19198580086231232]
  5719. receiving
  5720. ##### Sending to neptune:  active_relus :  0.476803139382 , 9613262.34 #####
  5721. ##### Sending to neptune:  dp_per_s :  0.476803139382 , 88.6668697671 #####
  5722. [u'other', 9613262.34, 88.66686976706502]
  5723. receiving
  5724. [0130 17:58:37 @multigpu.py:323] [p0574]  step: count(1191), step_time 1487.49, mean_step_time 1438.53, it/s 0.7
  5725. [0130 17:58:38 @multigpu.py:323] [p0115]  step: count(1177), step_time 1448.55, mean_step_time 1431.54, it/s 0.7
  5726. [0130 17:58:39 @multigpu.py:323] [p0576]  step: count(1201), step_time 1473.19, mean_step_time 1437.25, it/s 0.7
  5727. [0130 17:58:39 @multigpu.py:323] [p0574]  step: count(1192), step_time 1479.83, mean_step_time 1441.54, it/s 0.69
  5728. [0130 17:58:40 @multigpu.py:323] [p0115]  step: count(1178), step_time 1453.8, mean_step_time 1434.01, it/s 0.7
  5729. [0130 17:58:40 @multigpu.py:323] [p0574]  step: count(1193), step_time 1379.85, mean_step_time 1438.31, it/s 0.7
  5730. [0130 17:58:40 @multigpu.py:323] [p0576]  step: count(1202), step_time 1401.42, mean_step_time 1435.07, it/s 0.7
  5731. [0130 17:58:41 @multigpu.py:323] [p0115]  step: count(1179), step_time 1398.57, mean_step_time 1432.0, it/s 0.7
  5732. [0130 17:58:41 @multigpu.py:323] [p0574]  step: count(1194), step_time 1400.66, mean_step_time 1435.43, it/s 0.7
  5733. [0130 17:58:41 @multigpu.py:323] [p0576]  step: count(1203), step_time 1400.03, mean_step_time 1434.6, it/s 0.7
  5734. [0130 17:58:43 @multigpu.py:323] [p0576]  step: count(1204), step_time 1419.88, mean_step_time 1435.32, it/s 0.7
  5735. [0130 17:58:43 @multigpu.py:323] [p0574]  step: count(1195), step_time 1470.39, mean_step_time 1437.69, it/s 0.7
  5736. [0130 17:58:43 @multigpu.py:323] [p0115]  step: count(1180), step_time 1514.03, mean_step_time 1435.55, it/s 0.7
  5737. [0130 17:58:44 @multigpu.py:323] [p0574]  step: count(1196), step_time 1378.91, mean_step_time 1434.38, it/s 0.7
  5738. [0130 17:58:44 @multigpu.py:323] [p0576]  step: count(1205), step_time 1442.53, mean_step_time 1438.99, it/s 0.69
  5739. [0130 17:58:44 @multigpu.py:323] [p0115]  step: count(1181), step_time 1418.57, mean_step_time 1435.02, it/s 0.7
  5740. [0130 17:58:46 @multigpu.py:323] [p0574]  step: count(1197), step_time 1424.19, mean_step_time 1434.92, it/s 0.7
  5741. [0130 17:58:46 @multigpu.py:323] [p0576]  step: count(1206), step_time 1445.85, mean_step_time 1439.75, it/s 0.69
  5742. [0130 17:58:46 @multigpu.py:323] [p0115]  step: count(1182), step_time 1449.86, mean_step_time 1435.82, it/s 0.7
  5743. [0130 17:58:47 @multigpu.py:323] [p0576]  step: count(1207), step_time 1386.01, mean_step_time 1440.8, it/s 0.69
  5744. [0130 17:58:47 @multigpu.py:323] [p0574]  step: count(1198), step_time 1435.24, mean_step_time 1430.63, it/s 0.7
  5745. [0130 17:58:47 @multigpu.py:323] [p0115]  step: count(1183), step_time 1491.59, mean_step_time 1437.84, it/s 0.7
  5746. [0130 17:58:49 @multigpu.py:323] [p0576]  step: count(1208), step_time 1456.2, mean_step_time 1440.8, it/s 0.69
  5747. [0130 17:58:49 @multigpu.py:323] [p0574]  step: count(1199), step_time 1449.49, mean_step_time 1432.82, it/s 0.7
  5748. [0130 17:58:49 @multigpu.py:323] [p0115]  step: count(1184), step_time 1419.96, mean_step_time 1438.24, it/s 0.7
  5749. [0130 17:58:50 @multigpu.py:323] [p0576]  step: count(1209), step_time 1403.29, mean_step_time 1442.07, it/s 0.69
  5750. [0130 17:58:50 @multigpu.py:323] [p0574]  step: count(1200), step_time 1416.98, mean_step_time 1432.38, it/s 0.7
  5751. sending debugging info...
  5752. sending to address tcp://p0112:61216
  5753. ##### Sending to neptune:  mean_delay :  0.480370954143 , 0.0 #####
  5754. ##### Sending to neptune:  max_delay :  0.480370954143 , -0.0 #####
  5755. sending to address tcp://p0112:61216
  5756. ##### Sending to neptune:  min_delay :  0.480370954143 , -0.0 #####
  5757. [u'delays', [0.0, -0.0, -0.0]]
  5758. receiving
  5759. ##### Sending to neptune:  cost :  0.480371489988 , -0.00687376549467 #####
  5760. sending to address tcp://p0112:61216
  5761. ##### Sending to neptune:  policy_loss :  0.480371489988 , -0.0515709109604 #####
  5762. ##### Sending to neptune:  xentropy_loss :  0.480371489988 , -2.28922891617 #####
  5763. ##### Sending to neptune:  value_loss :  0.480371489988 , 1.46095776558 #####
  5764. ##### Sending to neptune:  advantage :  0.480371489988 , 0.000297629856504 #####
  5765. ##### Sending to neptune:  pred_reward :  0.480371489988 , 0.45008713007 #####
  5766. ##### Sending to neptune:  max_logit :  0.480371489988 , 0.190380632877 #####
  5767. [u'loss', -0.006873765494674444, -0.05157091096043587, -2.289228916168213, 1.4609577655792236, 0.0002976298565044999, 0.45008713006973267, 0.19038063287734985]
  5768. receiving
  5769. ##### Sending to neptune:  active_relus :  0.480371931063 , 9601806.97 #####
  5770. ##### Sending to neptune:  dp_per_s :  0.480371931063 , 88.6732990532 #####
  5771. [u'other', 9601806.97, 88.6732990532418]
  5772. receiving
  5773. [0130 17:58:50 @multigpu.py:323] [p0115]  step: count(1185), step_time 1475.1, mean_step_time 1442.45, it/s 0.69
  5774. [0130 17:58:51 @multigpu.py:323] [p0574]  step: count(1201), step_time 1401.01, mean_step_time 1431.56, it/s 0.7
  5775. [0130 17:58:51 @multigpu.py:323] [p0576]  step: count(1210), step_time 1431.24, mean_step_time 1442.46, it/s 0.69
  5776. [0130 17:58:52 @multigpu.py:323] [p0115]  step: count(1186), step_time 1473.77, mean_step_time 1444.5, it/s 0.69
  5777. [0130 17:58:53 @multigpu.py:323] [p0574]  step: count(1202), step_time 1393.14, mean_step_time 1426.42, it/s 0.7
  5778. [0130 17:58:53 @multigpu.py:323] [p0576]  step: count(1211), step_time 1452.95, mean_step_time 1444.43, it/s 0.69
  5779. [0130 17:58:53 @multigpu.py:323] [p0115]  step: count(1187), step_time 1401.6, mean_step_time 1443.15, it/s 0.69
  5780. [0130 17:58:54 @multigpu.py:323] [p0115]  step: count(1188), step_time 1399.15, mean_step_time 1435.26, it/s 0.7
  5781. [0130 17:58:54 @multigpu.py:323] [p0574]  step: count(1203), step_time 1659.28, mean_step_time 1435.29, it/s 0.7
  5782. [0130 17:58:54 @multigpu.py:323] [p0576]  step: count(1212), step_time 1584.36, mean_step_time 1435.26, it/s 0.7
  5783. [0130 17:58:56 @multigpu.py:323] [p0576]  step: count(1213), step_time 1403.86, mean_step_time 1432.09, it/s 0.7
  5784. sending to address tcp://p0112:61216
  5785. ##### Sending to neptune:  online_score :  0.482001303567 , 2.1 #####
  5786. [u'online', 2.1]
  5787. receiving
  5788. [0130 17:58:56 @multigpu.py:323] [p0115]  step: count(1189), step_time 1445.36, mean_step_time 1436.11, it/s 0.7
  5789. [0130 17:58:56 @multigpu.py:323] [p0574]  step: count(1204), step_time 1458.93, mean_step_time 1438.33, it/s 0.7
  5790. [0130 17:58:57 @multigpu.py:323] [p0574]  step: count(1205), step_time 1395.93, mean_step_time 1436.12, it/s 0.7
  5791. [0130 17:58:57 @multigpu.py:323] [p0576]  step: count(1214), step_time 1477.0, mean_step_time 1437.07, it/s 0.7
  5792. [0130 17:58:57 @multigpu.py:323] [p0115]  step: count(1190), step_time 1476.45, mean_step_time 1438.23, it/s 0.7
  5793. [0130 17:58:59 @multigpu.py:323] [p0574]  step: count(1206), step_time 1423.39, mean_step_time 1437.74, it/s 0.7
  5794. [0130 17:58:59 @multigpu.py:323] [p0576]  step: count(1215), step_time 1412.4, mean_step_time 1434.46, it/s 0.7
  5795. [0130 17:58:59 @multigpu.py:323] [p0115]  step: count(1191), step_time 1499.11, mean_step_time 1440.53, it/s 0.69
  5796. sending to address tcp://p0112:61216
  5797. ##### Sending to neptune:  online_score :  0.482869036661 , 2.0 #####
  5798. [u'online', 2.0]
  5799. receiving
  5800. [0130 17:59:00 @multigpu.py:323] [p0576]  step: count(1216), step_time 1419.01, mean_step_time 1435.37, it/s 0.7
  5801. [0130 17:59:00 @multigpu.py:323] [p0574]  step: count(1207), step_time 1436.29, mean_step_time 1438.62, it/s 0.7
  5802. [0130 17:59:00 @multigpu.py:323] [p0115]  step: count(1192), step_time 1446.43, mean_step_time 1443.06, it/s 0.69
  5803. [0130 17:59:02 @multigpu.py:323] [p0576]  step: count(1217), step_time 1414.54, mean_step_time 1435.16, it/s 0.7
  5804. [0130 17:59:02 @multigpu.py:323] [p0574]  step: count(1208), step_time 1452.89, mean_step_time 1439.09, it/s 0.69
  5805. [0130 17:59:02 @multigpu.py:323] [p0115]  step: count(1193), step_time 1457.53, mean_step_time 1444.29, it/s 0.69
  5806. [0130 17:59:03 @multigpu.py:323] [p0576]  step: count(1218), step_time 1407.46, mean_step_time 1432.84, it/s 0.7
  5807. [0130 17:59:03 @multigpu.py:323] [p0574]  step: count(1209), step_time 1390.8, mean_step_time 1438.22, it/s 0.7
  5808. [0130 17:59:03 @multigpu.py:323] [p0115]  step: count(1194), step_time 1429.94, mean_step_time 1446.71, it/s 0.69
  5809. [0130 17:59:04 @multigpu.py:323] [p0576]  step: count(1219), step_time 1411.38, mean_step_time 1431.3, it/s 0.7
  5810. [0130 17:59:04 @multigpu.py:323] [p0574]  step: count(1210), step_time 1437.77, mean_step_time 1438.62, it/s 0.7
  5811. [0130 17:59:05 @multigpu.py:323] [p0115]  step: count(1195), step_time 1420.12, mean_step_time 1446.35, it/s 0.69
  5812. [0130 17:59:06 @multigpu.py:323] [p0576]  step: count(1220), step_time 1431.94, mean_step_time 1433.73, it/s 0.7
  5813. [0130 17:59:06 @multigpu.py:323] [p0574]  step: count(1211), step_time 1405.34, mean_step_time 1434.52, it/s 0.7
  5814. [0130 17:59:06 @multigpu.py:323] [p0115]  step: count(1196), step_time 1424.1, mean_step_time 1447.18, it/s 0.69
  5815. sending to address tcp://p0112:61216
  5816. ##### Sending to neptune:  online_score :  0.484892646074 , 2.2 #####
  5817. [u'online', 2.2]
  5818. receiving
  5819. [0130 17:59:07 @multigpu.py:323] [p0576]  step: count(1221), step_time 1406.25, mean_step_time 1430.38, it/s 0.7
  5820. [0130 17:59:07 @multigpu.py:323] [p0574]  step: count(1212), step_time 1464.63, mean_step_time 1433.76, it/s 0.7
  5821. [0130 17:59:07 @multigpu.py:323] [p0115]  step: count(1197), step_time 1451.57, mean_step_time 1447.33, it/s 0.69
  5822. [0130 17:59:09 @multigpu.py:323] [p0576]  step: count(1222), step_time 1469.92, mean_step_time 1433.8, it/s 0.7
  5823. [0130 17:59:09 @multigpu.py:323] [p0574]  step: count(1213), step_time 1403.7, mean_step_time 1434.95, it/s 0.7
  5824. [0130 17:59:09 @multigpu.py:323] [p0115]  step: count(1198), step_time 1456.32, mean_step_time 1447.46, it/s 0.69
  5825. [0130 17:59:10 @multigpu.py:323] [p0576]  step: count(1223), step_time 1411.45, mean_step_time 1434.38, it/s 0.7
  5826. [0130 17:59:10 @multigpu.py:323] [p0574]  step: count(1214), step_time 1427.55, mean_step_time 1436.29, it/s 0.7
  5827. [0130 17:59:10 @multigpu.py:323] [p0115]  step: count(1199), step_time 1462.95, mean_step_time 1450.67, it/s 0.69
  5828. [0130 17:59:11 @multigpu.py:323] [p0576]  step: count(1224), step_time 1425.05, mean_step_time 1434.63, it/s 0.7
  5829. [0130 17:59:12 @multigpu.py:323] [p0574]  step: count(1215), step_time 1438.4, mean_step_time 1434.69, it/s 0.7
  5830. [0130 17:59:12 @multigpu.py:323] [p0115]  step: count(1200), step_time 1445.15, mean_step_time 1447.23, it/s 0.69
  5831. sending debugging info...
  5832. sending to address tcp://p0112:61216
  5833. ##### Sending to neptune:  mean_delay :  0.486447773841 , 0.0 #####
  5834. sending to address tcp://p0112:61216
  5835. ##### Sending to neptune:  max_delay :  0.486447773841 , -0.0 #####
  5836. ##### Sending to neptune:  min_delay :  0.486447773841 , -0.0 #####
  5837. [u'delays', [0.0, -0.0, -0.0]]
  5838. ##### Sending to neptune:  active_workers :  0.486448003054 , 3 #####
  5839. receiving
  5840. ##### Sending to neptune:  cost :  0.486448276374 , -0.00475755473599 #####
  5841. sending to address tcp://p0112:61216
  5842. ##### Sending to neptune:  policy_loss :  0.486448276374 , 0.0145185422152 #####
  5843. ##### Sending to neptune:  xentropy_loss :  0.486448276374 , -2.28904032707 #####
  5844. ##### Sending to neptune:  value_loss :  0.486448276374 , 1.66555476189 #####
  5845. ##### Sending to neptune:  advantage :  0.486448276374 , -9.42744300119e-05 #####
  5846. ##### Sending to neptune:  pred_reward :  0.486448276374 , 0.464911580086 #####
  5847. ##### Sending to neptune:  max_logit :  0.486448276374 , 0.191080719233 #####
  5848. [u'loss', -0.0047575547359883785, 0.01451854221522808, -2.2890403270721436, 1.6655547618865967, -9.427443001186475e-05, 0.4649115800857544, 0.1910807192325592]
  5849. receiving
  5850. ##### Sending to neptune:  active_relus :  0.486448774669 , 9607864.04 #####
  5851. ##### Sending to neptune:  dp_per_s :  0.486448774669 , 88.614272994 #####
  5852. [u'other', 9607864.04, 88.61427299401984]
  5853. receiving
  5854. [0130 17:59:13 @multigpu.py:323] [p0576]  step: count(1225), step_time 1419.94, mean_step_time 1433.5, it/s 0.7
  5855. [0130 17:59:13 @multigpu.py:323] [p0574]  step: count(1216), step_time 1427.25, mean_step_time 1437.11, it/s 0.7
  5856. [0130 17:59:13 @multigpu.py:323] [p0115]  step: count(1201), step_time 1423.0, mean_step_time 1447.45, it/s 0.69
  5857. [0130 17:59:14 @multigpu.py:323] [p0574]  step: count(1217), step_time 1435.25, mean_step_time 1437.66, it/s 0.7
  5858. [0130 17:59:14 @multigpu.py:323] [p0576]  step: count(1226), step_time 1497.93, mean_step_time 1436.11, it/s 0.7
  5859. [0130 17:59:15 @multigpu.py:323] [p0115]  step: count(1202), step_time 1460.86, mean_step_time 1448.0, it/s 0.69
  5860. [0130 17:59:16 @multigpu.py:323] [p0576]  step: count(1227), step_time 1387.57, mean_step_time 1436.19, it/s 0.7
  5861. [0130 17:59:16 @multigpu.py:323] [p0574]  step: count(1218), step_time 1428.02, mean_step_time 1437.3, it/s 0.7
  5862. [0130 17:59:16 @multigpu.py:323] [p0115]  step: count(1203), step_time 1384.71, mean_step_time 1442.66, it/s 0.69
  5863. [0130 17:59:17 @multigpu.py:323] [p0574]  step: count(1219), step_time 1419.66, mean_step_time 1435.81, it/s 0.7
  5864. [0130 17:59:17 @multigpu.py:323] [p0576]  step: count(1228), step_time 1456.31, mean_step_time 1436.19, it/s 0.7
  5865. [0130 17:59:18 @multigpu.py:323] [p0115]  step: count(1204), step_time 1445.15, mean_step_time 1443.92, it/s 0.69
  5866. [0130 17:59:19 @multigpu.py:323] [p0576]  step: count(1229), step_time 1376.86, mean_step_time 1434.87, it/s 0.7
  5867. [0130 17:59:19 @multigpu.py:323] [p0574]  step: count(1220), step_time 1418.2, mean_step_time 1435.87, it/s 0.7
  5868. [0130 17:59:19 @multigpu.py:323] [p0115]  step: count(1205), step_time 1404.72, mean_step_time 1440.4, it/s 0.69
  5869. [0130 17:59:20 @multigpu.py:323] [p0576]  step: count(1230), step_time 1425.41, mean_step_time 1434.58, it/s 0.7
  5870. sending to address tcp://p0112:61216
  5871. ##### Sending to neptune:  online_score :  0.488742454118 , 1.9 #####
  5872. [u'online', 1.9]
  5873. receiving
  5874. [0130 17:59:20 @multigpu.py:323] [p0574]  step: count(1221), step_time 1416.16, mean_step_time 1436.63, it/s 0.7
  5875. [0130 17:59:20 @multigpu.py:323] [p0115]  step: count(1206), step_time 1460.48, mean_step_time 1439.73, it/s 0.69
  5876. [0130 17:59:22 @multigpu.py:323] [p0576]  step: count(1231), step_time 1449.42, mean_step_time 1434.4, it/s 0.7
  5877. [0130 17:59:22 @multigpu.py:323] [p0574]  step: count(1222), step_time 1442.99, mean_step_time 1439.12, it/s 0.69
  5878. [0130 17:59:22 @multigpu.py:323] [p0115]  step: count(1207), step_time 1462.84, mean_step_time 1442.8, it/s 0.69
  5879. [0130 17:59:23 @multigpu.py:323] [p0115]  step: count(1208), step_time 1410.18, mean_step_time 1443.35, it/s 0.69
  5880. [0130 17:59:23 @multigpu.py:323] [p0576]  step: count(1232), step_time 1763.18, mean_step_time 1443.34, it/s 0.69
  5881. [0130 17:59:23 @multigpu.py:323] [p0574]  step: count(1223), step_time 1745.65, mean_step_time 1443.44, it/s 0.69
  5882. [0130 17:59:25 @multigpu.py:323] [p0576]  step: count(1233), step_time 1403.5, mean_step_time 1443.32, it/s 0.69
  5883. [0130 17:59:25 @multigpu.py:323] [p0115]  step: count(1209), step_time 1411.91, mean_step_time 1441.68, it/s 0.69
  5884. [0130 17:59:25 @multigpu.py:323] [p0574]  step: count(1224), step_time 1449.48, mean_step_time 1442.97, it/s 0.69
  5885. [0130 17:59:26 @multigpu.py:323] [p0576]  step: count(1234), step_time 1413.02, mean_step_time 1440.13, it/s 0.69
  5886. [0130 17:59:26 @multigpu.py:323] [p0115]  step: count(1210), step_time 1415.36, mean_step_time 1438.62, it/s 0.7
  5887. [0130 17:59:26 @multigpu.py:323] [p0574]  step: count(1225), step_time 1408.73, mean_step_time 1443.61, it/s 0.69
  5888. [0130 17:59:28 @multigpu.py:323] [p0576]  step: count(1235), step_time 1410.4, mean_step_time 1440.03, it/s 0.69
  5889. [0130 17:59:28 @multigpu.py:323] [p0574]  step: count(1226), step_time 1409.94, mean_step_time 1442.94, it/s 0.69
  5890. [0130 17:59:28 @multigpu.py:323] [p0115]  step: count(1211), step_time 1486.26, mean_step_time 1437.98, it/s 0.7
  5891. [0130 17:59:29 @multigpu.py:323] [p0576]  step: count(1236), step_time 1377.55, mean_step_time 1437.95, it/s 0.7
  5892. [0130 17:59:29 @multigpu.py:323] [p0574]  step: count(1227), step_time 1451.71, mean_step_time 1443.71, it/s 0.69
  5893. [0130 17:59:29 @multigpu.py:323] [p0115]  step: count(1212), step_time 1450.74, mean_step_time 1438.19, it/s 0.7
  5894. [0130 17:59:30 @multigpu.py:323] [p0576]  step: count(1237), step_time 1406.86, mean_step_time 1437.57, it/s 0.7
  5895. [0130 17:59:30 @multigpu.py:323] [p0574]  step: count(1228), step_time 1483.1, mean_step_time 1445.22, it/s 0.69
  5896. [0130 17:59:30 @multigpu.py:323] [p0115]  step: count(1213), step_time 1449.25, mean_step_time 1437.78, it/s 0.7
  5897. sending to address tcp://p0112:61216
  5898. ##### Sending to neptune:  online_score :  0.491776681079 , 1.5 #####
  5899. [u'online', 1.5]
  5900. receiving
  5901. [0130 17:59:32 @multigpu.py:323] [p0576]  step: count(1238), step_time 1422.62, mean_step_time 1438.33, it/s 0.7
  5902. [0130 17:59:32 @multigpu.py:323] [p0115]  step: count(1214), step_time 1383.4, mean_step_time 1435.45, it/s 0.7
  5903. [0130 17:59:32 @multigpu.py:323] [p0574]  step: count(1229), step_time 1452.75, mean_step_time 1448.31, it/s 0.69
  5904. [0130 17:59:33 @multigpu.py:323] [p0576]  step: count(1239), step_time 1400.1, mean_step_time 1437.76, it/s 0.7
  5905. [0130 17:59:33 @multigpu.py:323] [p0115]  step: count(1215), step_time 1418.15, mean_step_time 1435.36, it/s 0.7
  5906. [0130 17:59:33 @multigpu.py:323] [p0574]  step: count(1230), step_time 1462.29, mean_step_time 1449.54, it/s 0.69
  5907. [0130 17:59:35 @multigpu.py:323] [p0576]  step: count(1240), step_time 1443.06, mean_step_time 1438.32, it/s 0.7
  5908. [0130 17:59:35 @multigpu.py:323] [p0115]  step: count(1216), step_time 1425.73, mean_step_time 1435.44, it/s 0.7
  5909. [0130 17:59:35 @multigpu.py:323] [p0574]  step: count(1231), step_time 1409.05, mean_step_time 1449.73, it/s 0.69
  5910. [0130 17:59:36 @multigpu.py:323] [p0576]  step: count(1241), step_time 1385.11, mean_step_time 1437.26, it/s 0.7
  5911. [0130 17:59:36 @multigpu.py:323] [p0115]  step: count(1217), step_time 1434.85, mean_step_time 1434.6, it/s 0.7
  5912. [0130 17:59:36 @multigpu.py:323] [p0574]  step: count(1232), step_time 1401.94, mean_step_time 1446.59, it/s 0.69
  5913. [0130 17:59:37 @multigpu.py:323] [p0576]  step: count(1242), step_time 1437.12, mean_step_time 1435.62, it/s 0.7
  5914. [0130 17:59:38 @multigpu.py:323] [p0115]  step: count(1218), step_time 1427.21, mean_step_time 1433.15, it/s 0.7
  5915. [0130 17:59:38 @multigpu.py:323] [p0574]  step: count(1233), step_time 1444.78, mean_step_time 1448.65, it/s 0.69
  5916. sending to address tcp://p0112:61216
  5917. ##### Sending to neptune:  online_score :  0.493770758841 , 2.2 #####
  5918. [u'online', 2.2]
  5919. receiving
  5920. [0130 17:59:39 @multigpu.py:323] [p0576]  step: count(1243), step_time 1458.66, mean_step_time 1437.98, it/s 0.7
  5921. [0130 17:59:39 @multigpu.py:323] [p0115]  step: count(1219), step_time 1430.12, mean_step_time 1431.5, it/s 0.7
  5922. [0130 17:59:39 @multigpu.py:323] [p0574]  step: count(1234), step_time 1417.29, mean_step_time 1448.13, it/s 0.69
  5923. [0130 17:59:40 @multigpu.py:323] [p0576]  step: count(1244), step_time 1432.94, mean_step_time 1438.38, it/s 0.7
  5924. [0130 17:59:40 @multigpu.py:323] [p0115]  step: count(1220), step_time 1428.74, mean_step_time 1430.68, it/s 0.7
  5925. [0130 17:59:41 @multigpu.py:323] [p0574]  step: count(1235), step_time 1480.59, mean_step_time 1450.24, it/s 0.69
  5926. [0130 17:59:42 @multigpu.py:323] [p0576]  step: count(1245), step_time 1429.8, mean_step_time 1438.87, it/s 0.69
  5927. [0130 17:59:42 @multigpu.py:323] [p0115]  step: count(1221), step_time 1496.29, mean_step_time 1434.35, it/s 0.7
  5928. [0130 17:59:42 @multigpu.py:323] [p0574]  step: count(1236), step_time 1433.37, mean_step_time 1450.55, it/s 0.69
  5929. [0130 17:59:43 @multigpu.py:323] [p0576]  step: count(1246), step_time 1430.66, mean_step_time 1435.51, it/s 0.7
  5930. [0130 17:59:43 @multigpu.py:323] [p0115]  step: count(1222), step_time 1429.09, mean_step_time 1432.76, it/s 0.7
  5931. [0130 17:59:43 @multigpu.py:323] [p0574]  step: count(1237), step_time 1417.41, mean_step_time 1449.66, it/s 0.69
  5932. [0130 17:59:45 @multigpu.py:323] [p0576]  step: count(1247), step_time 1401.68, mean_step_time 1436.21, it/s 0.7
  5933. [0130 17:59:45 @multigpu.py:323] [p0115]  step: count(1223), step_time 1466.01, mean_step_time 1436.82, it/s 0.7
  5934. [0130 17:59:45 @multigpu.py:323] [p0574]  step: count(1238), step_time 1439.89, mean_step_time 1450.25, it/s 0.69
  5935. sending to address tcp://p0112:61216
  5936. ##### Sending to neptune:  online_score :  0.495688630276 , 1.6 #####
  5937. [u'online', 1.6]
  5938. receiving
  5939. [0130 17:59:46 @multigpu.py:323] [p0576]  step: count(1248), step_time 1470.4, mean_step_time 1436.92, it/s 0.7
  5940. [0130 17:59:46 @multigpu.py:323] [p0574]  step: count(1239), step_time 1392.69, mean_step_time 1448.9, it/s 0.69
  5941. [0130 17:59:46 @multigpu.py:323] [p0115]  step: count(1224), step_time 1429.94, mean_step_time 1436.06, it/s 0.7
  5942. [0130 17:59:47 @multigpu.py:323] [p0576]  step: count(1249), step_time 1418.38, mean_step_time 1438.99, it/s 0.69
  5943. [0130 17:59:48 @multigpu.py:323] [p0574]  step: count(1240), step_time 1410.31, mean_step_time 1448.51, it/s 0.69
  5944. [0130 17:59:48 @multigpu.py:323] [p0115]  step: count(1225), step_time 1454.57, mean_step_time 1438.56, it/s 0.7
  5945. [0130 17:59:49 @multigpu.py:323] [p0576]  step: count(1250), step_time 1469.11, mean_step_time 1441.18, it/s 0.69
  5946. [0130 17:59:49 @multigpu.py:323] [p0574]  step: count(1241), step_time 1455.97, mean_step_time 1450.5, it/s 0.69
  5947. [0130 17:59:49 @multigpu.py:323] [p0115]  step: count(1226), step_time 1404.38, mean_step_time 1435.75, it/s 0.7
  5948. [0130 17:59:50 @multigpu.py:323] [p0576]  step: count(1251), step_time 1474.91, mean_step_time 1442.45, it/s 0.69
  5949. [0130 17:59:50 @multigpu.py:323] [p0574]  step: count(1242), step_time 1392.37, mean_step_time 1447.97, it/s 0.69
  5950. [0130 17:59:51 @multigpu.py:323] [p0115]  step: count(1227), step_time 1441.44, mean_step_time 1434.68, it/s 0.7
  5951. [0130 17:59:52 @multigpu.py:323] [p0115]  step: count(1228), step_time 1428.7, mean_step_time 1435.61, it/s 0.7
  5952. [0130 17:59:52 @multigpu.py:323] [p0574]  step: count(1243), step_time 1497.43, mean_step_time 1435.55, it/s 0.7
  5953. [0130 17:59:52 @multigpu.py:323] [p0576]  step: count(1252), step_time 1625.91, mean_step_time 1435.59, it/s 0.7
  5954. [0130 17:59:53 @multigpu.py:323] [p0576]  step: count(1253), step_time 1398.41, mean_step_time 1435.33, it/s 0.7
  5955. [0130 17:59:53 @multigpu.py:323] [p0115]  step: count(1229), step_time 1408.68, mean_step_time 1435.45, it/s 0.7
  5956. [0130 17:59:53 @multigpu.py:323] [p0574]  step: count(1244), step_time 1411.42, mean_step_time 1433.65, it/s 0.7
  5957. [0130 17:59:55 @multigpu.py:323] [p0576]  step: count(1254), step_time 1420.87, mean_step_time 1435.73, it/s 0.7
  5958. [0130 17:59:55 @multigpu.py:323] [p0574]  step: count(1245), step_time 1447.2, mean_step_time 1435.57, it/s 0.7
  5959. [0130 17:59:55 @multigpu.py:323] [p0115]  step: count(1230), step_time 1484.63, mean_step_time 1438.91, it/s 0.69
  5960. [0130 17:59:56 @multigpu.py:323] [p0576]  step: count(1255), step_time 1397.69, mean_step_time 1435.09, it/s 0.7
  5961. [0130 17:59:56 @multigpu.py:323] [p0115]  step: count(1231), step_time 1391.17, mean_step_time 1434.15, it/s 0.7
  5962. [0130 17:59:56 @multigpu.py:323] [p0574]  step: count(1246), step_time 1455.19, mean_step_time 1437.84, it/s 0.7
  5963. [0130 17:59:58 @multigpu.py:323] [p0576]  step: count(1256), step_time 1380.83, mean_step_time 1435.26, it/s 0.7
  5964. [0130 17:59:58 @multigpu.py:323] [p0115]  step: count(1232), step_time 1437.02, mean_step_time 1433.47, it/s 0.7
  5965. [0130 17:59:58 @multigpu.py:323] [p0574]  step: count(1247), step_time 1427.1, mean_step_time 1436.61, it/s 0.7
  5966. [0130 17:59:59 @multigpu.py:323] [p0576]  step: count(1257), step_time 1488.55, mean_step_time 1439.34, it/s 0.69
  5967. [0130 17:59:59 @multigpu.py:323] [p0574]  step: count(1248), step_time 1399.27, mean_step_time 1432.42, it/s 0.7
  5968. [0130 17:59:59 @multigpu.py:323] [p0115]  step: count(1233), step_time 1424.49, mean_step_time 1432.23, it/s 0.7
  5969. [0130 18:00:01 @multigpu.py:323] [p0576]  step: count(1258), step_time 1418.93, mean_step_time 1439.15, it/s 0.69
  5970. [0130 18:00:01 @multigpu.py:323] [p0574]  step: count(1249), step_time 1382.59, mean_step_time 1428.91, it/s 0.7
  5971. [0130 18:00:01 @multigpu.py:323] [p0115]  step: count(1234), step_time 1432.11, mean_step_time 1434.67, it/s 0.7
  5972. sending to address tcp://p0112:61216
  5973. ##### Sending to neptune:  online_score :  0.500219986373 , 1.8 #####
  5974. [u'online', 1.8]
  5975. receiving
  5976. [0130 18:00:02 @multigpu.py:323] [p0576]  step: count(1259), step_time 1450.22, mean_step_time 1441.66, it/s 0.69
  5977. [0130 18:00:02 @multigpu.py:323] [p0115]  step: count(1235), step_time 1417.64, mean_step_time 1434.64, it/s 0.7
  5978. [0130 18:00:02 @multigpu.py:323] [p0574]  step: count(1250), step_time 1493.7, mean_step_time 1430.48, it/s 0.7
  5979. [0130 18:00:03 @multigpu.py:323] [p0576]  step: count(1260), step_time 1436.07, mean_step_time 1441.31, it/s 0.69
  5980. [0130 18:00:03 @multigpu.py:323] [p0115]  step: count(1236), step_time 1446.74, mean_step_time 1435.69, it/s 0.7
  5981. [0130 18:00:03 @multigpu.py:323] [p0574]  step: count(1251), step_time 1452.27, mean_step_time 1432.64, it/s 0.7
  5982. [0130 18:00:05 @multigpu.py:323] [p0576]  step: count(1261), step_time 1421.8, mean_step_time 1443.15, it/s 0.69
  5983. [0130 18:00:05 @multigpu.py:323] [p0574]  step: count(1252), step_time 1412.02, mean_step_time 1433.14, it/s 0.7
  5984. [0130 18:00:05 @multigpu.py:323] [p0115]  step: count(1237), step_time 1443.62, mean_step_time 1436.13, it/s 0.7
  5985. [0130 18:00:06 @multigpu.py:323] [p0115]  step: count(1238), step_time 1427.58, mean_step_time 1436.15, it/s 0.7
  5986. [0130 18:00:06 @multigpu.py:323] [p0576]  step: count(1262), step_time 1502.99, mean_step_time 1446.44, it/s 0.69
  5987. [0130 18:00:06 @multigpu.py:323] [p0574]  step: count(1253), step_time 1450.51, mean_step_time 1433.43, it/s 0.7
  5988. sending to address tcp://p0112:61216
  5989. ##### Sending to neptune:  online_score :  0.50163765053 , 1.8 #####
  5990. [u'online', 1.8]
  5991. receiving
  5992. [0130 18:00:08 @multigpu.py:323] [p0576]  step: count(1263), step_time 1376.02, mean_step_time 1442.31, it/s 0.69
  5993. [0130 18:00:08 @multigpu.py:323] [p0574]  step: count(1254), step_time 1377.62, mean_step_time 1431.45, it/s 0.7
  5994. [0130 18:00:08 @multigpu.py:323] [p0115]  step: count(1239), step_time 1444.37, mean_step_time 1436.86, it/s 0.7
  5995. [0130 18:00:09 @multigpu.py:323] [p0115]  step: count(1240), step_time 1384.45, mean_step_time 1434.65, it/s 0.7
  5996. [0130 18:00:09 @multigpu.py:323] [p0576]  step: count(1264), step_time 1465.98, mean_step_time 1443.96, it/s 0.69
  5997. [0130 18:00:09 @multigpu.py:323] [p0574]  step: count(1255), step_time 1491.25, mean_step_time 1431.98, it/s 0.7
  5998. [0130 18:00:11 @multigpu.py:323] [p0576]  step: count(1265), step_time 1377.57, mean_step_time 1441.35, it/s 0.69
  5999. [0130 18:00:11 @multigpu.py:323] [p0115]  step: count(1241), step_time 1427.5, mean_step_time 1431.21, it/s 0.7
  6000. [0130 18:00:11 @multigpu.py:323] [p0574]  step: count(1256), step_time 1425.77, mean_step_time 1431.6, it/s 0.7
  6001. [0130 18:00:12 @multigpu.py:323] [p0115]  step: count(1242), step_time 1413.5, mean_step_time 1430.43, it/s 0.7
  6002. [0130 18:00:12 @multigpu.py:323] [p0576]  step: count(1266), step_time 1485.86, mean_step_time 1444.11, it/s 0.69
  6003. [0130 18:00:12 @multigpu.py:323] [p0574]  step: count(1257), step_time 1398.03, mean_step_time 1430.63, it/s 0.7
  6004. [0130 18:00:13 @multigpu.py:323] [p0576]  step: count(1267), step_time 1384.89, mean_step_time 1443.27, it/s 0.69
  6005. [0130 18:00:13 @multigpu.py:323] [p0574]  step: count(1258), step_time 1404.02, mean_step_time 1428.84, it/s 0.7
  6006. [0130 18:00:13 @multigpu.py:323] [p0115]  step: count(1243), step_time 1449.9, mean_step_time 1429.62, it/s 0.7
  6007. [0130 18:00:15 @multigpu.py:323] [p0115]  step: count(1244), step_time 1399.48, mean_step_time 1428.1, it/s 0.7
  6008. [0130 18:00:15 @multigpu.py:323] [p0574]  step: count(1259), step_time 1415.75, mean_step_time 1429.99, it/s 0.7
  6009. [0130 18:00:15 @multigpu.py:323] [p0576]  step: count(1268), step_time 1455.89, mean_step_time 1442.54, it/s 0.69
  6010. [0130 18:00:16 @multigpu.py:323] [p0115]  step: count(1245), step_time 1399.33, mean_step_time 1425.34, it/s 0.7
  6011. [0130 18:00:16 @multigpu.py:323] [p0574]  step: count(1260), step_time 1419.75, mean_step_time 1430.46, it/s 0.7
  6012. [0130 18:00:16 @multigpu.py:323] [p0576]  step: count(1269), step_time 1447.95, mean_step_time 1444.02, it/s 0.69
  6013. sending to address tcp://p0112:61216
  6014. ##### Sending to neptune:  online_score :  0.504519854983 , 1.2 #####
  6015. [u'online', 1.2]
  6016. ##### Sending to neptune:  active_workers :  0.504519919422 , 3 #####
  6017. receiving
  6018. [0130 18:00:18 @multigpu.py:323] [p0115]  step: count(1246), step_time 1424.38, mean_step_time 1426.34, it/s 0.7
  6019. [0130 18:00:18 @multigpu.py:323] [p0574]  step: count(1261), step_time 1404.1, mean_step_time 1427.87, it/s 0.7
  6020. [0130 18:00:18 @multigpu.py:323] [p0576]  step: count(1270), step_time 1404.17, mean_step_time 1440.78, it/s 0.69
  6021. [0130 18:00:19 @multigpu.py:323] [p0574]  step: count(1262), step_time 1389.37, mean_step_time 1427.72, it/s 0.7
  6022. [0130 18:00:19 @multigpu.py:323] [p0115]  step: count(1247), step_time 1435.36, mean_step_time 1426.03, it/s 0.7
  6023. [0130 18:00:19 @multigpu.py:323] [p0576]  step: count(1271), step_time 1398.44, mean_step_time 1436.95, it/s 0.7
  6024. [0130 18:00:21 @multigpu.py:323] [p0115]  step: count(1248), step_time 1413.05, mean_step_time 1425.25, it/s 0.7
  6025. [0130 18:00:21 @multigpu.py:323] [p0576]  step: count(1272), step_time 1389.99, mean_step_time 1425.16, it/s 0.7
  6026. [0130 18:00:21 @multigpu.py:323] [p0574]  step: count(1263), step_time 1449.4, mean_step_time 1425.32, it/s 0.7
  6027. [0130 18:00:22 @multigpu.py:323] [p0576]  step: count(1273), step_time 1400.21, mean_step_time 1425.25, it/s 0.7
  6028. [0130 18:00:22 @multigpu.py:323] [p0574]  step: count(1264), step_time 1407.21, mean_step_time 1425.11, it/s 0.7
  6029. [0130 18:00:22 @multigpu.py:323] [p0115]  step: count(1249), step_time 1461.25, mean_step_time 1427.88, it/s 0.7
  6030. [0130 18:00:23 @multigpu.py:323] [p0576]  step: count(1274), step_time 1490.45, mean_step_time 1428.73, it/s 0.7
  6031. [0130 18:00:23 @multigpu.py:323] [p0115]  step: count(1250), step_time 1429.73, mean_step_time 1425.13, it/s 0.7
  6032. [0130 18:00:23 @multigpu.py:323] [p0574]  step: count(1265), step_time 1537.12, mean_step_time 1429.6, it/s 0.7
  6033. [0130 18:00:25 @multigpu.py:323] [p0576]  step: count(1275), step_time 1365.34, mean_step_time 1427.11, it/s 0.7
  6034. [0130 18:00:25 @multigpu.py:323] [p0115]  step: count(1251), step_time 1407.17, mean_step_time 1425.93, it/s 0.7
  6035. [0130 18:00:25 @multigpu.py:323] [p0574]  step: count(1266), step_time 1356.73, mean_step_time 1424.68, it/s 0.7
  6036. [0130 18:00:26 @multigpu.py:323] [p0576]  step: count(1276), step_time 1412.74, mean_step_time 1428.7, it/s 0.7
  6037. [0130 18:00:26 @multigpu.py:323] [p0115]  step: count(1252), step_time 1420.38, mean_step_time 1425.1, it/s 0.7
  6038. [0130 18:00:26 @multigpu.py:323] [p0574]  step: count(1267), step_time 1491.01, mean_step_time 1427.87, it/s 0.7
  6039. [0130 18:00:28 @multigpu.py:323] [p0576]  step: count(1277), step_time 1476.89, mean_step_time 1428.12, it/s 0.7
  6040. [0130 18:00:28 @multigpu.py:323] [p0115]  step: count(1253), step_time 1436.83, mean_step_time 1425.72, it/s 0.7
  6041. [0130 18:00:28 @multigpu.py:323] [p0574]  step: count(1268), step_time 1375.61, mean_step_time 1426.69, it/s 0.7
  6042. [0130 18:00:29 @multigpu.py:323] [p0576]  step: count(1278), step_time 1414.23, mean_step_time 1427.89, it/s 0.7
  6043. [0130 18:00:29 @multigpu.py:323] [p0115]  step: count(1254), step_time 1456.16, mean_step_time 1426.92, it/s 0.7
  6044. [0130 18:00:29 @multigpu.py:323] [p0574]  step: count(1269), step_time 1470.79, mean_step_time 1431.1, it/s 0.7
  6045. [0130 18:00:31 @multigpu.py:323] [p0576]  step: count(1279), step_time 1439.79, mean_step_time 1427.36, it/s 0.7
  6046. [0130 18:00:31 @multigpu.py:323] [p0115]  step: count(1255), step_time 1390.55, mean_step_time 1425.57, it/s 0.7
  6047. [0130 18:00:31 @multigpu.py:323] [p0574]  step: count(1270), step_time 1458.49, mean_step_time 1429.34, it/s 0.7
  6048. [0130 18:00:32 @multigpu.py:323] [p0576]  step: count(1280), step_time 1430.71, mean_step_time 1427.1, it/s 0.7
  6049. [0130 18:00:32 @multigpu.py:323] [p0115]  step: count(1256), step_time 1428.18, mean_step_time 1424.64, it/s 0.7
  6050. [0130 18:00:32 @multigpu.py:323] [p0574]  step: count(1271), step_time 1435.24, mean_step_time 1428.49, it/s 0.7
  6051. [0130 18:00:33 @multigpu.py:323] [p0115]  step: count(1257), step_time 1449.35, mean_step_time 1424.92, it/s 0.7
  6052. [0130 18:00:33 @multigpu.py:323] [p0576]  step: count(1281), step_time 1465.3, mean_step_time 1429.27, it/s 0.7
  6053. [0130 18:00:33 @multigpu.py:323] [p0574]  step: count(1272), step_time 1448.46, mean_step_time 1430.31, it/s 0.7
  6054. sending to address tcp://p0112:61216
  6055. ##### Sending to neptune:  online_score :  0.509209632476 , 1.7 #####
  6056. [u'online', 1.7]
  6057. receiving
  6058. sending to address tcp://p0112:61216
  6059. ##### Sending to neptune:  online_score :  0.509221385519 , 1.8 #####
  6060. [u'online', 1.8]
  6061. receiving
  6062. [0130 18:00:35 @multigpu.py:323] [p0576]  step: count(1282), step_time 1407.69, mean_step_time 1424.51, it/s 0.7
  6063. [0130 18:00:35 @multigpu.py:323] [p0115]  step: count(1258), step_time 1431.65, mean_step_time 1425.13, it/s 0.7
  6064. [0130 18:00:35 @multigpu.py:323] [p0574]  step: count(1273), step_time 1424.04, mean_step_time 1428.99, it/s 0.7
  6065. [0130 18:00:36 @multigpu.py:323] [p0115]  step: count(1259), step_time 1441.86, mean_step_time 1425.0, it/s 0.7
  6066. [0130 18:00:36 @multigpu.py:323] [p0576]  step: count(1283), step_time 1467.47, mean_step_time 1429.08, it/s 0.7
  6067. [0130 18:00:36 @multigpu.py:323] [p0574]  step: count(1274), step_time 1414.32, mean_step_time 1430.82, it/s 0.7
  6068. [0130 18:00:38 @multigpu.py:323] [p0115]  step: count(1260), step_time 1417.55, mean_step_time 1426.66, it/s 0.7
  6069. [0130 18:00:38 @multigpu.py:323] [p0576]  step: count(1284), step_time 1448.14, mean_step_time 1428.19, it/s 0.7
  6070. [0130 18:00:38 @multigpu.py:323] [p0574]  step: count(1275), step_time 1415.54, mean_step_time 1427.04, it/s 0.7
  6071. sending to address tcp://p0112:61216
  6072. ##### Sending to neptune:  online_score :  0.510377633837 , 1.3 #####
  6073. [u'online', 1.3]
  6074. receiving
  6075. [0130 18:00:39 @multigpu.py:323] [p0115]  step: count(1261), step_time 1374.91, mean_step_time 1424.03, it/s 0.7
  6076. [0130 18:00:39 @multigpu.py:323] [p0576]  step: count(1285), step_time 1424.08, mean_step_time 1430.51, it/s 0.7
  6077. [0130 18:00:39 @multigpu.py:323] [p0574]  step: count(1276), step_time 1443.27, mean_step_time 1427.91, it/s 0.7
  6078. [0130 18:00:40 @multigpu.py:323] [p0115]  step: count(1262), step_time 1429.56, mean_step_time 1424.83, it/s 0.7
  6079. [0130 18:00:41 @multigpu.py:323] [p0576]  step: count(1286), step_time 1391.27, mean_step_time 1425.78, it/s 0.7
  6080. [0130 18:00:41 @multigpu.py:323] [p0574]  step: count(1277), step_time 1455.26, mean_step_time 1430.77, it/s 0.7
  6081. [0130 18:00:42 @multigpu.py:323] [p0115]  step: count(1263), step_time 1416.86, mean_step_time 1423.18, it/s 0.7
  6082. [0130 18:00:42 @multigpu.py:323] [p0576]  step: count(1287), step_time 1418.46, mean_step_time 1427.46, it/s 0.7
  6083. [0130 18:00:42 @multigpu.py:323] [p0574]  step: count(1278), step_time 1389.05, mean_step_time 1430.03, it/s 0.7
  6084. [0130 18:00:43 @multigpu.py:323] [p0115]  step: count(1264), step_time 1434.03, mean_step_time 1424.91, it/s 0.7
  6085. [0130 18:00:43 @multigpu.py:323] [p0576]  step: count(1288), step_time 1383.01, mean_step_time 1423.82, it/s 0.7
  6086. [0130 18:00:44 @multigpu.py:323] [p0574]  step: count(1279), step_time 1482.24, mean_step_time 1433.35, it/s 0.7
  6087. [0130 18:00:45 @multigpu.py:323] [p0576]  step: count(1289), step_time 1412.39, mean_step_time 1422.04, it/s 0.7
  6088. [0130 18:00:45 @multigpu.py:323] [p0115]  step: count(1265), step_time 1457.3, mean_step_time 1427.8, it/s 0.7
  6089. [0130 18:00:45 @multigpu.py:323] [p0574]  step: count(1280), step_time 1369.54, mean_step_time 1430.84, it/s 0.7
  6090. [0130 18:00:46 @multigpu.py:323] [p0576]  step: count(1290), step_time 1454.01, mean_step_time 1424.53, it/s 0.7
  6091. [0130 18:00:46 @multigpu.py:323] [p0115]  step: count(1266), step_time 1466.19, mean_step_time 1429.9, it/s 0.7
  6092. [0130 18:00:46 @multigpu.py:323] [p0574]  step: count(1281), step_time 1429.17, mean_step_time 1432.09, it/s 0.7
  6093. [0130 18:00:48 @multigpu.py:323] [p0576]  step: count(1291), step_time 1417.98, mean_step_time 1425.51, it/s 0.7
  6094. [0130 18:00:48 @multigpu.py:323] [p0115]  step: count(1267), step_time 1467.76, mean_step_time 1431.52, it/s 0.7
  6095. [0130 18:00:48 @multigpu.py:323] [p0574]  step: count(1282), step_time 1453.63, mean_step_time 1435.31, it/s 0.7
  6096. [0130 18:00:49 @multigpu.py:323] [p0115]  step: count(1268), step_time 1453.14, mean_step_time 1433.52, it/s 0.7
  6097. [0130 18:00:49 @multigpu.py:323] [p0574]  step: count(1283), step_time 1408.98, mean_step_time 1433.28, it/s 0.7
  6098. [0130 18:00:49 @multigpu.py:323] [p0576]  step: count(1292), step_time 1550.01, mean_step_time 1433.51, it/s 0.7
  6099. [0130 18:00:51 @multigpu.py:323] [p0576]  step: count(1293), step_time 1398.16, mean_step_time 1433.41, it/s 0.7
  6100. [0130 18:00:51 @multigpu.py:323] [p0115]  step: count(1269), step_time 1414.4, mean_step_time 1431.18, it/s 0.7
  6101. [0130 18:00:51 @multigpu.py:323] [p0574]  step: count(1284), step_time 1436.26, mean_step_time 1434.74, it/s 0.7
  6102. [0130 18:00:52 @multigpu.py:323] [p0115]  step: count(1270), step_time 1387.23, mean_step_time 1429.05, it/s 0.7
  6103. [0130 18:00:52 @multigpu.py:323] [p0576]  step: count(1294), step_time 1448.65, mean_step_time 1431.32, it/s 0.7
  6104. [0130 18:00:52 @multigpu.py:323] [p0574]  step: count(1285), step_time 1440.68, mean_step_time 1429.91, it/s 0.7
  6105. [0130 18:00:53 @multigpu.py:323] [p0115]  step: count(1271), step_time 1413.52, mean_step_time 1429.37, it/s 0.7
  6106. [0130 18:00:53 @multigpu.py:323] [p0574]  step: count(1286), step_time 1391.43, mean_step_time 1431.65, it/s 0.7
  6107. [0130 18:00:53 @multigpu.py:323] [p0576]  step: count(1295), step_time 1450.11, mean_step_time 1435.56, it/s 0.7
  6108. [0130 18:00:55 @multigpu.py:323] [p0115]  step: count(1272), step_time 1425.58, mean_step_time 1429.63, it/s 0.7
  6109. [0130 18:00:55 @multigpu.py:323] [p0574]  step: count(1287), step_time 1399.37, mean_step_time 1427.07, it/s 0.7
  6110. [0130 18:00:55 @multigpu.py:323] [p0576]  step: count(1296), step_time 1421.49, mean_step_time 1435.99, it/s 0.7
  6111. [0130 18:00:56 @multigpu.py:323] [p0115]  step: count(1273), step_time 1414.14, mean_step_time 1428.5, it/s 0.7
  6112. [0130 18:00:56 @multigpu.py:323] [p0574]  step: count(1288), step_time 1441.05, mean_step_time 1430.34, it/s 0.7
  6113. [0130 18:00:56 @multigpu.py:323] [p0576]  step: count(1297), step_time 1401.6, mean_step_time 1432.23, it/s 0.7
  6114. sending to address tcp://p0112:61216
  6115. ##### Sending to neptune:  online_score :  0.515617745519 , 1.5 #####
  6116. [u'online', 1.5]
  6117. receiving
  6118. [0130 18:00:58 @multigpu.py:323] [p0115]  step: count(1274), step_time 1429.4, mean_step_time 1427.16, it/s 0.7
  6119. [0130 18:00:58 @multigpu.py:323] [p0576]  step: count(1298), step_time 1386.47, mean_step_time 1430.84, it/s 0.7
  6120. [0130 18:00:58 @multigpu.py:323] [p0574]  step: count(1289), step_time 1408.34, mean_step_time 1427.22, it/s 0.7
  6121. [0130 18:00:59 @multigpu.py:323] [p0576]  step: count(1299), step_time 1403.87, mean_step_time 1429.04, it/s 0.7
  6122. [0130 18:00:59 @multigpu.py:323] [p0115]  step: count(1275), step_time 1463.58, mean_step_time 1430.81, it/s 0.7
  6123. [0130 18:00:59 @multigpu.py:323] [p0574]  step: count(1290), step_time 1480.56, mean_step_time 1428.32, it/s 0.7
  6124. [0130 18:01:01 @multigpu.py:323] [p0576]  step: count(1300), step_time 1430.43, mean_step_time 1429.03, it/s 0.7
  6125. sending debugging info...
  6126. sending to address tcp://p0112:61216
  6127. ##### Sending to neptune:  mean_delay :  0.516647246612 , 0.0 #####
  6128. sending to address tcp://p0112:61216
  6129. ##### Sending to neptune:  max_delay :  0.516647246612 , -0.0 #####
  6130. ##### Sending to neptune:  min_delay :  0.516647246612 , -0.0 #####
  6131. [u'delays', [0.0, -0.0, -0.0]]
  6132. receiving
  6133. ##### Sending to neptune:  cost :  0.51664776716 , -0.00946542713791 #####
  6134. sending to address tcp://p0112:61216
  6135. ##### Sending to neptune:  policy_loss :  0.51664776716 , -0.313773274422 #####
  6136. ##### Sending to neptune:  xentropy_loss :  0.51664776716 , -2.28776264191 #####
  6137. ##### Sending to neptune:  value_loss :  0.51664776716 , 1.38996088505 #####
  6138. ##### Sending to neptune:  advantage :  0.51664776716 , 0.00140806438867 #####
  6139. ##### Sending to neptune:  pred_reward :  0.51664776716 , 0.47912633419 #####
  6140. ##### Sending to neptune:  max_logit :  0.51664776716 , 0.194705918431 #####
  6141. [u'loss', -0.00946542713791132, -0.3137732744216919, -2.2877626419067383, 1.3899608850479126, 0.0014080643886700273, 0.47912633419036865, 0.19470591843128204]
  6142. receiving
  6143. ##### Sending to neptune:  active_relus :  0.516648219956 , 9676168.09 #####
  6144. ##### Sending to neptune:  dp_per_s :  0.516648219956 , 89.1823527596 #####
  6145. [u'other', 9676168.09, 89.18235275962654]
  6146. receiving
  6147. [0130 18:01:01 @multigpu.py:323] [p0574]  step: count(1291), step_time 1398.52, mean_step_time 1426.49, it/s 0.7
  6148. [0130 18:01:01 @multigpu.py:323] [p0115]  step: count(1276), step_time 1465.75, mean_step_time 1432.69, it/s 0.7
  6149. [0130 18:01:02 @multigpu.py:323] [p0576]  step: count(1301), step_time 1438.25, mean_step_time 1427.68, it/s 0.7
  6150. [0130 18:01:02 @multigpu.py:323] [p0574]  step: count(1292), step_time 1438.86, mean_step_time 1426.0, it/s 0.7
  6151. [0130 18:01:02 @multigpu.py:323] [p0115]  step: count(1277), step_time 1457.61, mean_step_time 1433.1, it/s 0.7
  6152. [0130 18:01:03 @multigpu.py:323] [p0576]  step: count(1302), step_time 1384.81, mean_step_time 1426.53, it/s 0.7
  6153. [0130 18:01:03 @multigpu.py:323] [p0115]  step: count(1278), step_time 1413.94, mean_step_time 1432.22, it/s 0.7
  6154. [0130 18:01:03 @multigpu.py:323] [p0574]  step: count(1293), step_time 1452.43, mean_step_time 1427.42, it/s 0.7
  6155. [0130 18:01:05 @multigpu.py:323] [p0576]  step: count(1303), step_time 1399.88, mean_step_time 1423.15, it/s 0.7
  6156. [0130 18:01:05 @multigpu.py:323] [p0574]  step: count(1294), step_time 1376.51, mean_step_time 1425.53, it/s 0.7
  6157. [0130 18:01:05 @multigpu.py:323] [p0115]  step: count(1279), step_time 1481.14, mean_step_time 1434.18, it/s 0.7
  6158. [0130 18:01:06 @multigpu.py:323] [p0576]  step: count(1304), step_time 1466.55, mean_step_time 1424.07, it/s 0.7
  6159. [0130 18:01:06 @multigpu.py:323] [p0574]  step: count(1295), step_time 1445.89, mean_step_time 1427.05, it/s 0.7
  6160. [0130 18:01:06 @multigpu.py:323] [p0115]  step: count(1280), step_time 1414.5, mean_step_time 1434.03, it/s 0.7
  6161. sending to address tcp://p0112:61216
  6162. ##### Sending to neptune:  online_score :  0.518410489957 , 2.2 #####
  6163. [u'online', 2.2]
  6164. receiving
  6165. [0130 18:01:08 @multigpu.py:323] [p0576]  step: count(1305), step_time 1397.55, mean_step_time 1422.75, it/s 0.7
  6166. [0130 18:01:08 @multigpu.py:323] [p0574]  step: count(1296), step_time 1387.43, mean_step_time 1424.26, it/s 0.7
  6167. [0130 18:01:08 @multigpu.py:323] [p0115]  step: count(1281), step_time 1412.38, mean_step_time 1435.9, it/s 0.7
  6168. [0130 18:01:09 @multigpu.py:323] [p0576]  step: count(1306), step_time 1436.32, mean_step_time 1425.0, it/s 0.7
  6169. [0130 18:01:09 @multigpu.py:323] [p0574]  step: count(1297), step_time 1474.04, mean_step_time 1425.2, it/s 0.7
  6170. [0130 18:01:09 @multigpu.py:323] [p0115]  step: count(1282), step_time 1413.31, mean_step_time 1435.09, it/s 0.7
  6171. [0130 18:01:10 @multigpu.py:323] [p0576]  step: count(1307), step_time 1420.68, mean_step_time 1425.11, it/s 0.7
  6172. [0130 18:01:11 @multigpu.py:323] [p0115]  step: count(1283), step_time 1401.14, mean_step_time 1434.3, it/s 0.7
  6173. [0130 18:01:11 @multigpu.py:323] [p0574]  step: count(1298), step_time 1436.18, mean_step_time 1427.56, it/s 0.7
  6174. [0130 18:01:12 @multigpu.py:323] [p0576]  step: count(1308), step_time 1451.23, mean_step_time 1428.52, it/s 0.7
  6175. [0130 18:01:12 @multigpu.py:323] [p0574]  step: count(1299), step_time 1454.11, mean_step_time 1426.15, it/s 0.7
  6176. [0130 18:01:12 @multigpu.py:323] [p0115]  step: count(1284), step_time 1456.81, mean_step_time 1435.44, it/s 0.7
  6177. [0130 18:01:13 @multigpu.py:323] [p0576]  step: count(1309), step_time 1384.38, mean_step_time 1427.12, it/s 0.7
  6178. [0130 18:01:13 @multigpu.py:323] [p0115]  step: count(1285), step_time 1412.14, mean_step_time 1433.18, it/s 0.7
  6179. [0130 18:01:14 @multigpu.py:323] [p0574]  step: count(1300), step_time 1449.74, mean_step_time 1430.16, it/s 0.7
  6180. sending debugging info...
  6181. sending to address tcp://p0112:61216
  6182. ##### Sending to neptune:  mean_delay :  0.520251732734 , 0.0 #####
  6183. sending to address tcp://p0112:61216
  6184. ##### Sending to neptune:  max_delay :  0.520251732734 , -0.0 #####
  6185. ##### Sending to neptune:  min_delay :  0.520251732734 , -0.0 #####
  6186. [u'delays', [0.0, -0.0, -0.0]]
  6187. receiving
  6188. ##### Sending to neptune:  cost :  0.520252199107 , -0.00133042002562 #####
  6189. sending to address tcp://p0112:61216
  6190. ##### Sending to neptune:  policy_loss :  0.520252199107 , 0.530481159687 #####
  6191. ##### Sending to neptune:  xentropy_loss :  0.520252199107 , -2.28754925728 #####
  6192. ##### Sending to neptune:  value_loss :  0.520252199107 , 1.58677446842 #####
  6193. ##### Sending to neptune:  advantage :  0.520252199107 , -0.0023368424736 #####
  6194. ##### Sending to neptune:  pred_reward :  0.520252199107 , 0.480544269085 #####
  6195. ##### Sending to neptune:  max_logit :  0.520252199107 , 0.196291729808 #####
  6196. [u'loss', -0.0013304200256243348, 0.5304811596870422, -2.2875492572784424, 1.586774468421936, -0.0023368424735963345, 0.4805442690849304, 0.1962917298078537]
  6197. receiving
  6198. ##### Sending to neptune:  active_relus :  0.52025263634 , 9689908.7 #####
  6199. ##### Sending to neptune:  dp_per_s :  0.52025263634 , 89.2147275754 #####
  6200. [u'other', 9689908.7, 89.21472757536068]
  6201. receiving
  6202. sending to address tcp://p0112:61216
  6203. ##### Sending to neptune:  online_score :  0.520315144393 , 1.5 #####
  6204. [u'online', 1.5]
  6205. receiving
  6206. [0130 18:01:15 @multigpu.py:323] [p0576]  step: count(1310), step_time 1453.13, mean_step_time 1427.08, it/s 0.7
  6207. [0130 18:01:15 @multigpu.py:323] [p0115]  step: count(1286), step_time 1374.66, mean_step_time 1428.61, it/s 0.7
  6208. [0130 18:01:15 @multigpu.py:323] [p0574]  step: count(1301), step_time 1403.41, mean_step_time 1428.87, it/s 0.7
  6209. [0130 18:01:16 @multigpu.py:323] [p0576]  step: count(1311), step_time 1410.4, mean_step_time 1426.7, it/s 0.7
  6210. [0130 18:01:16 @multigpu.py:323] [p0574]  step: count(1302), step_time 1379.33, mean_step_time 1425.16, it/s 0.7
  6211. [0130 18:01:16 @multigpu.py:323] [p0115]  step: count(1287), step_time 1459.62, mean_step_time 1428.2, it/s 0.7
  6212. [0130 18:01:18 @multigpu.py:323] [p0115]  step: count(1288), step_time 1434.8, mean_step_time 1427.28, it/s 0.7
  6213. [0130 18:01:18 @multigpu.py:323] [p0574]  step: count(1303), step_time 1451.91, mean_step_time 1427.3, it/s 0.7
  6214. [0130 18:01:18 @multigpu.py:323] [p0576]  step: count(1312), step_time 1563.34, mean_step_time 1427.37, it/s 0.7
  6215. [0130 18:01:19 @multigpu.py:323] [p0115]  step: count(1289), step_time 1395.34, mean_step_time 1426.33, it/s 0.7
  6216. [0130 18:01:19 @multigpu.py:323] [p0576]  step: count(1313), step_time 1409.53, mean_step_time 1427.93, it/s 0.7
  6217. [0130 18:01:19 @multigpu.py:323] [p0574]  step: count(1304), step_time 1418.34, mean_step_time 1426.41, it/s 0.7
  6218. [0130 18:01:21 @multigpu.py:323] [p0574]  step: count(1305), step_time 1398.31, mean_step_time 1424.29, it/s 0.7
  6219. [0130 18:01:21 @multigpu.py:323] [p0576]  step: count(1314), step_time 1409.09, mean_step_time 1425.96, it/s 0.7
  6220. [0130 18:01:21 @multigpu.py:323] [p0115]  step: count(1290), step_time 1429.8, mean_step_time 1428.46, it/s 0.7
  6221. [0130 18:01:22 @multigpu.py:323] [p0576]  step: count(1315), step_time 1402.83, mean_step_time 1423.59, it/s 0.7
  6222. [0130 18:01:22 @multigpu.py:323] [p0115]  step: count(1291), step_time 1396.94, mean_step_time 1427.63, it/s 0.7
  6223. [0130 18:01:22 @multigpu.py:323] [p0574]  step: count(1306), step_time 1445.26, mean_step_time 1426.98, it/s 0.7
  6224. [0130 18:01:23 @multigpu.py:323] [p0576]  step: count(1316), step_time 1414.48, mean_step_time 1423.24, it/s 0.7
  6225. [0130 18:01:23 @multigpu.py:323] [p0115]  step: count(1292), step_time 1442.41, mean_step_time 1428.47, it/s 0.7
  6226. [0130 18:01:23 @multigpu.py:323] [p0574]  step: count(1307), step_time 1436.64, mean_step_time 1428.84, it/s 0.7
  6227. [0130 18:01:25 @multigpu.py:323] [p0115]  step: count(1293), step_time 1459.15, mean_step_time 1430.72, it/s 0.7
  6228. [0130 18:01:25 @multigpu.py:323] [p0574]  step: count(1308), step_time 1428.54, mean_step_time 1428.22, it/s 0.7
  6229. [0130 18:01:25 @multigpu.py:323] [p0576]  step: count(1317), step_time 1506.03, mean_step_time 1428.46, it/s 0.7
  6230. [0130 18:01:26 @multigpu.py:323] [p0115]  step: count(1294), step_time 1388.66, mean_step_time 1428.68, it/s 0.7
  6231. [0130 18:01:26 @multigpu.py:323] [p0574]  step: count(1309), step_time 1408.21, mean_step_time 1428.21, it/s 0.7
  6232. [0130 18:01:26 @multigpu.py:323] [p0576]  step: count(1318), step_time 1477.43, mean_step_time 1433.01, it/s 0.7
  6233. sending to address tcp://p0112:61216
  6234. ##### Sending to neptune:  online_score :  0.523968138827 , 0.8 #####
  6235. [u'online', 0.8]
  6236. ##### Sending to neptune:  active_workers :  0.523968183862 , 3 #####
  6237. receiving
  6238. [0130 18:01:28 @multigpu.py:323] [p0574]  step: count(1310), step_time 1415.15, mean_step_time 1424.94, it/s 0.7
  6239. [0130 18:01:28 @multigpu.py:323] [p0115]  step: count(1295), step_time 1442.54, mean_step_time 1427.63, it/s 0.7
  6240. [0130 18:01:28 @multigpu.py:323] [p0576]  step: count(1319), step_time 1366.87, mean_step_time 1431.16, it/s 0.7
  6241. [0130 18:01:29 @multigpu.py:323] [p0115]  step: count(1296), step_time 1352.47, mean_step_time 1421.97, it/s 0.7
  6242. [0130 18:01:29 @multigpu.py:323] [p0576]  step: count(1320), step_time 1396.82, mean_step_time 1429.48, it/s 0.7
  6243. [0130 18:01:29 @multigpu.py:323] [p0574]  step: count(1311), step_time 1455.72, mean_step_time 1427.8, it/s 0.7
  6244. [0130 18:01:31 @multigpu.py:323] [p0115]  step: count(1297), step_time 1460.09, mean_step_time 1422.09, it/s 0.7
  6245. [0130 18:01:31 @multigpu.py:323] [p0576]  step: count(1321), step_time 1402.61, mean_step_time 1427.7, it/s 0.7
  6246. [0130 18:01:31 @multigpu.py:323] [p0574]  step: count(1312), step_time 1426.71, mean_step_time 1427.19, it/s 0.7
  6247. [0130 18:01:32 @multigpu.py:323] [p0576]  step: count(1322), step_time 1393.86, mean_step_time 1428.15, it/s 0.7
  6248. [0130 18:01:32 @multigpu.py:323] [p0115]  step: count(1298), step_time 1479.78, mean_step_time 1425.38, it/s 0.7
  6249. [0130 18:01:32 @multigpu.py:323] [p0574]  step: count(1313), step_time 1424.25, mean_step_time 1425.79, it/s 0.7
  6250. sending to address tcp://p0112:61216
  6251. ##### Sending to neptune:  online_score :  0.525484263566 , 2.1 #####
  6252. [u'online', 2.1]
  6253. receiving
  6254. [0130 18:01:33 @multigpu.py:323] [p0576]  step: count(1323), step_time 1420.64, mean_step_time 1429.19, it/s 0.7
  6255. [0130 18:01:33 @multigpu.py:323] [p0115]  step: count(1299), step_time 1479.2, mean_step_time 1425.29, it/s 0.7
  6256. [0130 18:01:34 @multigpu.py:323] [p0574]  step: count(1314), step_time 1495.65, mean_step_time 1431.74, it/s 0.7
  6257. [0130 18:01:35 @multigpu.py:323] [p0576]  step: count(1324), step_time 1431.46, mean_step_time 1427.43, it/s 0.7
  6258. [0130 18:01:35 @multigpu.py:323] [p0574]  step: count(1315), step_time 1379.58, mean_step_time 1428.43, it/s 0.7
  6259. [0130 18:01:35 @multigpu.py:323] [p0115]  step: count(1300), step_time 1504.77, mean_step_time 1429.8, it/s 0.7
  6260. sending debugging info...
  6261. sending to address tcp://p0112:61216
  6262. ##### Sending to neptune:  mean_delay :  0.526216461658 , 0.0 #####
  6263. sending to address tcp://p0112:61216
  6264. ##### Sending to neptune:  max_delay :  0.526216461658 , -0.0 #####
  6265. ##### Sending to neptune:  min_delay :  0.526216461658 , -0.0 #####
  6266. [u'delays', [0.0, -0.0, -0.0]]
  6267. receiving
  6268. ##### Sending to neptune:  cost :  0.526216974987 , -0.0132103003561 #####
  6269. sending to address tcp://p0112:61216
  6270. ##### Sending to neptune:  policy_loss :  0.526216974987 , -0.767839074135 #####
  6271. ##### Sending to neptune:  xentropy_loss :  0.526216974987 , -2.2872877121 #####
  6272. ##### Sending to neptune:  value_loss :  0.526216974987 , 1.36420881748 #####
  6273. ##### Sending to neptune:  advantage :  0.526216974987 , 0.00328848976642 #####
  6274. ##### Sending to neptune:  pred_reward :  0.526216974987 , 0.47394412756 #####
  6275. ##### Sending to neptune:  max_logit :  0.526216974987 , 0.197944834828 #####
  6276. [u'loss', -0.013210300356149673, -0.7678390741348267, -2.287287712097168, 1.3642088174819946, 0.003288489766418934, 0.47394412755966187, 0.19794483482837677]
  6277. receiving
  6278. ##### Sending to neptune:  active_relus :  0.526217454142 , 9719758.83 #####
  6279. ##### Sending to neptune:  dp_per_s :  0.526217454142 , 89.3935747297 #####
  6280. [u'other', 9719758.83, 89.39357472974253]
  6281. receiving
  6282. [0130 18:01:36 @multigpu.py:323] [p0576]  step: count(1325), step_time 1412.64, mean_step_time 1428.19, it/s 0.7
  6283. [0130 18:01:36 @multigpu.py:323] [p0574]  step: count(1316), step_time 1420.66, mean_step_time 1430.09, it/s 0.7
  6284. [0130 18:01:36 @multigpu.py:323] [p0115]  step: count(1301), step_time 1411.45, mean_step_time 1429.75, it/s 0.7
  6285. [0130 18:01:38 @multigpu.py:323] [p0576]  step: count(1326), step_time 1386.73, mean_step_time 1425.71, it/s 0.7
  6286. [0130 18:01:38 @multigpu.py:323] [p0574]  step: count(1317), step_time 1391.52, mean_step_time 1425.96, it/s 0.7
  6287. [0130 18:01:38 @multigpu.py:323] [p0115]  step: count(1302), step_time 1399.21, mean_step_time 1429.05, it/s 0.7
  6288. sending to address tcp://p0112:61216
  6289. ##### Sending to neptune:  online_score :  0.527047368553 , 2.0 #####
  6290. [u'online', 2.0]
  6291. receiving
  6292. [0130 18:01:39 @multigpu.py:323] [p0576]  step: count(1327), step_time 1457.68, mean_step_time 1427.56, it/s 0.7
  6293. [0130 18:01:39 @multigpu.py:323] [p0574]  step: count(1318), step_time 1424.61, mean_step_time 1425.38, it/s 0.7
  6294. [0130 18:01:39 @multigpu.py:323] [p0115]  step: count(1303), step_time 1447.28, mean_step_time 1431.36, it/s 0.7
  6295. [0130 18:01:41 @multigpu.py:323] [p0576]  step: count(1328), step_time 1488.34, mean_step_time 1429.41, it/s 0.7
  6296. [0130 18:01:41 @multigpu.py:323] [p0574]  step: count(1319), step_time 1429.17, mean_step_time 1424.14, it/s 0.7
  6297. [0130 18:01:41 @multigpu.py:323] [p0115]  step: count(1304), step_time 1399.1, mean_step_time 1428.47, it/s 0.7
  6298. [0130 18:01:42 @multigpu.py:323] [p0576]  step: count(1329), step_time 1403.8, mean_step_time 1430.39, it/s 0.7
  6299. [0130 18:01:42 @multigpu.py:323] [p0574]  step: count(1320), step_time 1445.3, mean_step_time 1423.91, it/s 0.7
  6300. [0130 18:01:42 @multigpu.py:323] [p0115]  step: count(1305), step_time 1497.48, mean_step_time 1432.74, it/s 0.7
  6301. [0130 18:01:43 @multigpu.py:323] [p0576]  step: count(1330), step_time 1451.89, mean_step_time 1430.32, it/s 0.7
  6302. [0130 18:01:43 @multigpu.py:323] [p0574]  step: count(1321), step_time 1398.76, mean_step_time 1423.68, it/s 0.7
  6303. [0130 18:01:44 @multigpu.py:323] [p0115]  step: count(1306), step_time 1401.08, mean_step_time 1434.06, it/s 0.7
  6304. [0130 18:01:45 @multigpu.py:323] [p0576]  step: count(1331), step_time 1440.01, mean_step_time 1431.8, it/s 0.7
  6305. [0130 18:01:45 @multigpu.py:323] [p0574]  step: count(1322), step_time 1454.99, mean_step_time 1427.46, it/s 0.7
  6306. [0130 18:01:45 @multigpu.py:323] [p0115]  step: count(1307), step_time 1458.27, mean_step_time 1433.99, it/s 0.7
  6307. [0130 18:01:46 @multigpu.py:323] [p0574]  step: count(1323), step_time 1604.47, mean_step_time 1435.09, it/s 0.7
  6308. [0130 18:01:46 @multigpu.py:323] [p0576]  step: count(1332), step_time 1627.95, mean_step_time 1435.03, it/s 0.7
  6309. [0130 18:01:46 @multigpu.py:323] [p0115]  step: count(1308), step_time 1458.09, mean_step_time 1435.16, it/s 0.7
  6310. [0130 18:01:48 @multigpu.py:323] [p0576]  step: count(1333), step_time 1431.92, mean_step_time 1436.15, it/s 0.7
  6311. [0130 18:01:48 @multigpu.py:323] [p0574]  step: count(1324), step_time 1443.86, mean_step_time 1436.37, it/s 0.7
  6312. [0130 18:01:48 @multigpu.py:323] [p0115]  step: count(1309), step_time 1452.55, mean_step_time 1438.02, it/s 0.7
  6313. [0130 18:01:49 @multigpu.py:323] [p0115]  step: count(1310), step_time 1416.31, mean_step_time 1437.34, it/s 0.7
  6314. [0130 18:01:49 @multigpu.py:323] [p0576]  step: count(1334), step_time 1437.23, mean_step_time 1437.56, it/s 0.7
  6315. [0130 18:01:49 @multigpu.py:323] [p0574]  step: count(1325), step_time 1444.67, mean_step_time 1438.69, it/s 0.7
  6316. [0130 18:01:51 @multigpu.py:323] [p0574]  step: count(1326), step_time 1445.52, mean_step_time 1438.7, it/s 0.7
  6317. [0130 18:01:51 @multigpu.py:323] [p0576]  step: count(1335), step_time 1479.94, mean_step_time 1441.42, it/s 0.69
  6318. [0130 18:01:51 @multigpu.py:323] [p0115]  step: count(1311), step_time 1513.3, mean_step_time 1443.16, it/s 0.69
  6319. sending to address tcp://p0112:61216
  6320. ##### Sending to neptune:  online_score :  0.530835908585 , 1.7 #####
  6321. [u'online', 1.7]
  6322. receiving
  6323. [0130 18:01:52 @multigpu.py:323] [p0574]  step: count(1327), step_time 1439.03, mean_step_time 1438.82, it/s 0.7
  6324. [0130 18:01:52 @multigpu.py:323] [p0115]  step: count(1312), step_time 1434.38, mean_step_time 1442.76, it/s 0.69
  6325. [0130 18:01:52 @multigpu.py:323] [p0576]  step: count(1336), step_time 1471.0, mean_step_time 1444.24, it/s 0.69
  6326. [0130 18:01:54 @multigpu.py:323] [p0574]  step: count(1328), step_time 1425.5, mean_step_time 1438.67, it/s 0.7
  6327. [0130 18:01:54 @multigpu.py:323] [p0115]  step: count(1313), step_time 1428.68, mean_step_time 1441.23, it/s 0.69
  6328. [0130 18:01:54 @multigpu.py:323] [p0576]  step: count(1337), step_time 1450.34, mean_step_time 1441.46, it/s 0.69
  6329. [0130 18:01:55 @multigpu.py:323] [p0574]  step: count(1329), step_time 1386.17, mean_step_time 1437.56, it/s 0.7
  6330. [0130 18:01:55 @multigpu.py:323] [p0115]  step: count(1314), step_time 1403.13, mean_step_time 1441.96, it/s 0.69
  6331. [0130 18:01:55 @multigpu.py:323] [p0576]  step: count(1338), step_time 1405.76, mean_step_time 1437.87, it/s 0.7
  6332. [0130 18:01:57 @multigpu.py:323] [p0574]  step: count(1330), step_time 1465.31, mean_step_time 1440.07, it/s 0.69
  6333. [0130 18:01:57 @multigpu.py:323] [p0576]  step: count(1339), step_time 1444.95, mean_step_time 1441.78, it/s 0.69
  6334. [0130 18:01:57 @multigpu.py:323] [p0115]  step: count(1315), step_time 1514.84, mean_step_time 1445.57, it/s 0.69
  6335. [0130 18:01:58 @multigpu.py:323] [p0574]  step: count(1331), step_time 1451.43, mean_step_time 1439.86, it/s 0.69
  6336. [0130 18:01:58 @multigpu.py:323] [p0576]  step: count(1340), step_time 1451.64, mean_step_time 1444.52, it/s 0.69
  6337. [0130 18:01:58 @multigpu.py:323] [p0115]  step: count(1316), step_time 1430.77, mean_step_time 1449.49, it/s 0.69
  6338. [0130 18:01:59 @multigpu.py:323] [p0574]  step: count(1332), step_time 1405.66, mean_step_time 1438.8, it/s 0.7
  6339. [0130 18:01:59 @multigpu.py:323] [p0576]  step: count(1341), step_time 1411.03, mean_step_time 1444.94, it/s 0.69
  6340. [0130 18:02:00 @multigpu.py:323] [p0115]  step: count(1317), step_time 1455.08, mean_step_time 1449.24, it/s 0.69
  6341. [0130 18:02:01 @multigpu.py:323] [p0574]  step: count(1333), step_time 1401.36, mean_step_time 1437.66, it/s 0.7
  6342. [0130 18:02:01 @multigpu.py:323] [p0576]  step: count(1342), step_time 1452.52, mean_step_time 1447.87, it/s 0.69
  6343. [0130 18:02:01 @multigpu.py:323] [p0115]  step: count(1318), step_time 1443.73, mean_step_time 1447.43, it/s 0.69
  6344. [0130 18:02:02 @multigpu.py:323] [p0574]  step: count(1334), step_time 1450.79, mean_step_time 1435.42, it/s 0.7
  6345. [0130 18:02:02 @multigpu.py:323] [p0576]  step: count(1343), step_time 1401.61, mean_step_time 1446.92, it/s 0.69
  6346. [0130 18:02:02 @multigpu.py:323] [p0115]  step: count(1319), step_time 1411.88, mean_step_time 1444.07, it/s 0.69
  6347. sending to address tcp://p0112:61216
  6348. ##### Sending to neptune:  online_score :  0.533943954706 , 1.5 #####
  6349. [u'online', 1.5]
  6350. receiving
  6351. sending to address tcp://p0112:61216
  6352. ##### Sending to neptune:  online_score :  0.533990806341 , 1.0 #####
  6353. [u'online', 1.0]
  6354. receiving
  6355. [0130 18:02:04 @multigpu.py:323] [p0574]  step: count(1335), step_time 1388.52, mean_step_time 1435.86, it/s 0.7
  6356. [0130 18:02:04 @multigpu.py:323] [p0115]  step: count(1320), step_time 1406.04, mean_step_time 1439.13, it/s 0.69
  6357. [0130 18:02:04 @multigpu.py:323] [p0576]  step: count(1344), step_time 1489.33, mean_step_time 1449.82, it/s 0.69
  6358. [0130 18:02:05 @multigpu.py:323] [p0574]  step: count(1336), step_time 1413.78, mean_step_time 1435.52, it/s 0.7
  6359. [0130 18:02:05 @multigpu.py:323] [p0115]  step: count(1321), step_time 1403.22, mean_step_time 1438.72, it/s 0.7
  6360. [0130 18:02:05 @multigpu.py:323] [p0576]  step: count(1345), step_time 1422.62, mean_step_time 1450.32, it/s 0.69
  6361. [0130 18:02:06 @multigpu.py:323] [p0574]  step: count(1337), step_time 1472.75, mean_step_time 1439.58, it/s 0.69
  6362. [0130 18:02:07 @multigpu.py:323] [p0115]  step: count(1322), step_time 1451.03, mean_step_time 1441.31, it/s 0.69
  6363. [0130 18:02:07 @multigpu.py:323] [p0576]  step: count(1346), step_time 1488.87, mean_step_time 1455.42, it/s 0.69
  6364. [0130 18:02:08 @multigpu.py:323] [p0574]  step: count(1338), step_time 1503.68, mean_step_time 1443.54, it/s 0.69
  6365. [0130 18:02:08 @multigpu.py:323] [p0115]  step: count(1323), step_time 1434.57, mean_step_time 1440.68, it/s 0.69
  6366. [0130 18:02:08 @multigpu.py:323] [p0576]  step: count(1347), step_time 1404.22, mean_step_time 1452.75, it/s 0.69
  6367. [0130 18:02:09 @multigpu.py:323] [p0574]  step: count(1339), step_time 1424.48, mean_step_time 1443.3, it/s 0.69
  6368. [0130 18:02:09 @multigpu.py:323] [p0115]  step: count(1324), step_time 1374.08, mean_step_time 1439.42, it/s 0.69
  6369. [0130 18:02:10 @multigpu.py:323] [p0576]  step: count(1348), step_time 1401.08, mean_step_time 1448.39, it/s 0.69
  6370. [0130 18:02:11 @multigpu.py:323] [p0574]  step: count(1340), step_time 1442.77, mean_step_time 1443.18, it/s 0.69
  6371. [0130 18:02:11 @multigpu.py:323] [p0115]  step: count(1325), step_time 1459.87, mean_step_time 1437.54, it/s 0.7
  6372. [0130 18:02:11 @multigpu.py:323] [p0576]  step: count(1349), step_time 1440.85, mean_step_time 1450.24, it/s 0.69
  6373. [0130 18:02:12 @multigpu.py:323] [p0574]  step: count(1341), step_time 1431.65, mean_step_time 1444.82, it/s 0.69
  6374. [0130 18:02:12 @multigpu.py:323] [p0115]  step: count(1326), step_time 1427.27, mean_step_time 1438.85, it/s 0.69
  6375. [0130 18:02:12 @multigpu.py:323] [p0576]  step: count(1350), step_time 1395.23, mean_step_time 1447.41, it/s 0.69
  6376. [0130 18:02:14 @multigpu.py:323] [p0115]  step: count(1327), step_time 1408.36, mean_step_time 1436.36, it/s 0.7
  6377. [0130 18:02:14 @multigpu.py:323] [p0574]  step: count(1342), step_time 1472.65, mean_step_time 1445.7, it/s 0.69
  6378. [0130 18:02:14 @multigpu.py:323] [p0576]  step: count(1351), step_time 1438.57, mean_step_time 1447.33, it/s 0.69
  6379. [0130 18:02:15 @multigpu.py:323] [p0115]  step: count(1328), step_time 1482.76, mean_step_time 1437.59, it/s 0.7
  6380. [0130 18:02:15 @multigpu.py:323] [p0576]  step: count(1352), step_time 1433.1, mean_step_time 1437.59, it/s 0.7
  6381. [0130 18:02:15 @multigpu.py:323] [p0574]  step: count(1343), step_time 1443.32, mean_step_time 1437.65, it/s 0.7
  6382. [0130 18:02:17 @multigpu.py:323] [p0574]  step: count(1344), step_time 1387.03, mean_step_time 1434.8, it/s 0.7
  6383. [0130 18:02:17 @multigpu.py:323] [p0576]  step: count(1353), step_time 1433.47, mean_step_time 1437.67, it/s 0.7
  6384. [0130 18:02:17 @multigpu.py:323] [p0115]  step: count(1329), step_time 1459.3, mean_step_time 1437.93, it/s 0.7
  6385. [0130 18:02:18 @multigpu.py:323] [p0576]  step: count(1354), step_time 1409.46, mean_step_time 1436.28, it/s 0.7
  6386. [0130 18:02:18 @multigpu.py:323] [p0574]  step: count(1345), step_time 1457.71, mean_step_time 1435.46, it/s 0.7
  6387. [0130 18:02:18 @multigpu.py:323] [p0115]  step: count(1330), step_time 1426.61, mean_step_time 1438.44, it/s 0.7
  6388. [0130 18:02:19 @multigpu.py:323] [p0574]  step: count(1346), step_time 1415.38, mean_step_time 1433.95, it/s 0.7
  6389. [0130 18:02:20 @multigpu.py:323] [p0576]  step: count(1355), step_time 1492.05, mean_step_time 1436.89, it/s 0.7
  6390. [0130 18:02:20 @multigpu.py:323] [p0115]  step: count(1331), step_time 1460.89, mean_step_time 1435.82, it/s 0.7
  6391. [0130 18:02:21 @multigpu.py:323] [p0574]  step: count(1347), step_time 1426.43, mean_step_time 1433.32, it/s 0.7
  6392. [0130 18:02:21 @multigpu.py:323] [p0115]  step: count(1332), step_time 1375.86, mean_step_time 1432.9, it/s 0.7
  6393. [0130 18:02:21 @multigpu.py:323] [p0576]  step: count(1356), step_time 1408.51, mean_step_time 1433.76, it/s 0.7
  6394. [0130 18:02:22 @multigpu.py:323] [p0574]  step: count(1348), step_time 1386.07, mean_step_time 1431.35, it/s 0.7
  6395. [0130 18:02:22 @multigpu.py:323] [p0115]  step: count(1333), step_time 1430.46, mean_step_time 1432.99, it/s 0.7
  6396. [0130 18:02:22 @multigpu.py:323] [p0576]  step: count(1357), step_time 1415.37, mean_step_time 1432.01, it/s 0.7
  6397. [0130 18:02:24 @multigpu.py:323] [p0574]  step: count(1349), step_time 1406.08, mean_step_time 1432.34, it/s 0.7
  6398. [0130 18:02:24 @multigpu.py:323] [p0576]  step: count(1358), step_time 1453.69, mean_step_time 1434.41, it/s 0.7
  6399. [0130 18:02:24 @multigpu.py:323] [p0115]  step: count(1334), step_time 1488.36, mean_step_time 1437.25, it/s 0.7
  6400. [0130 18:02:25 @multigpu.py:323] [p0574]  step: count(1350), step_time 1441.79, mean_step_time 1431.17, it/s 0.7
  6401. [0130 18:02:25 @multigpu.py:323] [p0576]  step: count(1359), step_time 1387.2, mean_step_time 1431.52, it/s 0.7
  6402. [0130 18:02:25 @multigpu.py:323] [p0115]  step: count(1335), step_time 1428.06, mean_step_time 1432.91, it/s 0.7
  6403. [0130 18:02:27 @multigpu.py:323] [p0574]  step: count(1351), step_time 1414.6, mean_step_time 1429.33, it/s 0.7
  6404. [0130 18:02:27 @multigpu.py:323] [p0115]  step: count(1336), step_time 1423.21, mean_step_time 1432.53, it/s 0.7
  6405. [0130 18:02:27 @multigpu.py:323] [p0576]  step: count(1360), step_time 1503.81, mean_step_time 1434.13, it/s 0.7
  6406. sending to address tcp://p0112:61216
  6407. ##### Sending to neptune:  online_score :  0.540751250254 , 1.5 #####
  6408. [u'online', 1.5]
  6409. ##### Sending to neptune:  active_workers :  0.540751353833 , 3 #####
  6410. receiving
  6411. [0130 18:02:28 @multigpu.py:323] [p0574]  step: count(1352), step_time 1405.06, mean_step_time 1429.3, it/s 0.7
  6412. [0130 18:02:28 @multigpu.py:323] [p0576]  step: count(1361), step_time 1404.19, mean_step_time 1433.79, it/s 0.7
  6413. [0130 18:02:28 @multigpu.py:323] [p0115]  step: count(1337), step_time 1450.23, mean_step_time 1432.29, it/s 0.7
  6414. [0130 18:02:29 @multigpu.py:323] [p0574]  step: count(1353), step_time 1415.19, mean_step_time 1429.99, it/s 0.7
  6415. [0130 18:02:30 @multigpu.py:323] [p0576]  step: count(1362), step_time 1451.61, mean_step_time 1433.74, it/s 0.7
  6416. [0130 18:02:30 @multigpu.py:323] [p0115]  step: count(1338), step_time 1441.16, mean_step_time 1432.16, it/s 0.7
  6417. sending to address tcp://p0112:61216
  6418. ##### Sending to neptune:  online_score :  0.541542358332 , 1.7 #####
  6419. [u'online', 1.7]
  6420. receiving
  6421. [0130 18:02:31 @multigpu.py:323] [p0574]  step: count(1354), step_time 1390.2, mean_step_time 1426.96, it/s 0.7
  6422. [0130 18:02:31 @multigpu.py:323] [p0576]  step: count(1363), step_time 1408.87, mean_step_time 1434.11, it/s 0.7
  6423. [0130 18:02:31 @multigpu.py:323] [p0115]  step: count(1339), step_time 1413.46, mean_step_time 1432.24, it/s 0.7
  6424. [0130 18:02:32 @multigpu.py:323] [p0574]  step: count(1355), step_time 1440.46, mean_step_time 1429.55, it/s 0.7
  6425. [0130 18:02:32 @multigpu.py:323] [p0576]  step: count(1364), step_time 1434.43, mean_step_time 1431.36, it/s 0.7
  6426. [0130 18:02:33 @multigpu.py:323] [p0115]  step: count(1340), step_time 1494.65, mean_step_time 1436.67, it/s 0.7
  6427. sending to address tcp://p0112:61216
  6428. ##### Sending to neptune:  online_score :  0.542352651358 , 1.7 #####
  6429. [u'online', 1.7]
  6430. receiving
  6431. [0130 18:02:34 @multigpu.py:323] [p0574]  step: count(1356), step_time 1404.31, mean_step_time 1429.08, it/s 0.7
  6432. [0130 18:02:34 @multigpu.py:323] [p0576]  step: count(1365), step_time 1415.21, mean_step_time 1430.99, it/s 0.7
  6433. [0130 18:02:34 @multigpu.py:323] [p0115]  step: count(1341), step_time 1412.43, mean_step_time 1437.13, it/s 0.7
  6434. [0130 18:02:35 @multigpu.py:323] [p0574]  step: count(1357), step_time 1387.35, mean_step_time 1424.81, it/s 0.7
  6435. [0130 18:02:35 @multigpu.py:323] [p0576]  step: count(1366), step_time 1408.33, mean_step_time 1426.96, it/s 0.7
  6436. [0130 18:02:35 @multigpu.py:323] [p0115]  step: count(1342), step_time 1432.7, mean_step_time 1436.21, it/s 0.7
  6437. [0130 18:02:36 @multigpu.py:323] [p0574]  step: count(1358), step_time 1415.84, mean_step_time 1420.42, it/s 0.7
  6438. [0130 18:02:37 @multigpu.py:323] [p0576]  step: count(1367), step_time 1433.29, mean_step_time 1428.42, it/s 0.7
  6439. [0130 18:02:37 @multigpu.py:323] [p0115]  step: count(1343), step_time 1448.73, mean_step_time 1436.92, it/s 0.7
  6440. [0130 18:02:38 @multigpu.py:323] [p0574]  step: count(1359), step_time 1483.98, mean_step_time 1423.39, it/s 0.7
  6441. [0130 18:02:38 @multigpu.py:323] [p0576]  step: count(1368), step_time 1436.63, mean_step_time 1430.19, it/s 0.7
  6442. [0130 18:02:38 @multigpu.py:323] [p0115]  step: count(1344), step_time 1448.18, mean_step_time 1440.63, it/s 0.69
  6443. [0130 18:02:39 @multigpu.py:323] [p0574]  step: count(1360), step_time 1426.99, mean_step_time 1422.6, it/s 0.7
  6444. [0130 18:02:40 @multigpu.py:323] [p0576]  step: count(1369), step_time 1404.52, mean_step_time 1428.38, it/s 0.7
  6445. [0130 18:02:40 @multigpu.py:323] [p0115]  step: count(1345), step_time 1413.7, mean_step_time 1438.32, it/s 0.7
  6446. [0130 18:02:41 @multigpu.py:323] [p0574]  step: count(1361), step_time 1372.54, mean_step_time 1419.65, it/s 0.7
  6447. [0130 18:02:41 @multigpu.py:323] [p0576]  step: count(1370), step_time 1417.31, mean_step_time 1429.48, it/s 0.7
  6448. [0130 18:02:41 @multigpu.py:323] [p0115]  step: count(1346), step_time 1437.96, mean_step_time 1438.85, it/s 0.69
  6449. [0130 18:02:42 @multigpu.py:323] [p0574]  step: count(1362), step_time 1486.76, mean_step_time 1420.35, it/s 0.7
  6450. [0130 18:02:42 @multigpu.py:323] [p0576]  step: count(1371), step_time 1376.86, mean_step_time 1426.4, it/s 0.7
  6451. [0130 18:02:43 @multigpu.py:323] [p0115]  step: count(1347), step_time 1406.57, mean_step_time 1438.76, it/s 0.7
  6452. [0130 18:02:44 @multigpu.py:323] [p0115]  step: count(1348), step_time 1442.11, mean_step_time 1436.73, it/s 0.7
  6453. [0130 18:02:44 @multigpu.py:323] [p0576]  step: count(1372), step_time 1638.05, mean_step_time 1436.64, it/s 0.7
  6454. [0130 18:02:44 @multigpu.py:323] [p0574]  step: count(1363), step_time 1770.49, mean_step_time 1436.71, it/s 0.7
  6455. [0130 18:02:45 @multigpu.py:323] [p0576]  step: count(1373), step_time 1429.14, mean_step_time 1436.43, it/s 0.7
  6456. [0130 18:02:45 @multigpu.py:323] [p0574]  step: count(1364), step_time 1429.9, mean_step_time 1438.86, it/s 0.69
  6457. [0130 18:02:45 @multigpu.py:323] [p0115]  step: count(1349), step_time 1432.33, mean_step_time 1435.38, it/s 0.7
  6458. [0130 18:02:47 @multigpu.py:323] [p0576]  step: count(1374), step_time 1433.5, mean_step_time 1437.63, it/s 0.7
  6459. [0130 18:02:47 @multigpu.py:323] [p0574]  step: count(1365), step_time 1432.46, mean_step_time 1437.59, it/s 0.7
  6460. [0130 18:02:47 @multigpu.py:323] [p0115]  step: count(1350), step_time 1484.87, mean_step_time 1438.3, it/s 0.7
  6461. [0130 18:02:48 @multigpu.py:323] [p0574]  step: count(1366), step_time 1406.44, mean_step_time 1437.15, it/s 0.7
  6462. [0130 18:02:48 @multigpu.py:323] [p0576]  step: count(1375), step_time 1415.47, mean_step_time 1433.8, it/s 0.7
  6463. [0130 18:02:48 @multigpu.py:323] [p0115]  step: count(1351), step_time 1411.37, mean_step_time 1435.82, it/s 0.7
  6464. [0130 18:02:50 @multigpu.py:323] [p0574]  step: count(1367), step_time 1408.95, mean_step_time 1436.27, it/s 0.7
  6465. [0130 18:02:50 @multigpu.py:323] [p0576]  step: count(1376), step_time 1485.84, mean_step_time 1437.67, it/s 0.7
  6466. [0130 18:02:50 @multigpu.py:323] [p0115]  step: count(1352), step_time 1461.74, mean_step_time 1440.11, it/s 0.69
  6467. [0130 18:02:51 @multigpu.py:323] [p0574]  step: count(1368), step_time 1430.87, mean_step_time 1438.51, it/s 0.7
  6468. [0130 18:02:51 @multigpu.py:323] [p0115]  step: count(1353), step_time 1385.82, mean_step_time 1437.88, it/s 0.7
  6469. [0130 18:02:51 @multigpu.py:323] [p0576]  step: count(1377), step_time 1434.55, mean_step_time 1438.63, it/s 0.7
  6470. sending to address tcp://p0112:61216
  6471. ##### Sending to neptune:  online_score :  0.547551592191 , 1.8 #####
  6472. [u'online', 1.8]
  6473. receiving
  6474. [0130 18:02:52 @multigpu.py:323] [p0574]  step: count(1369), step_time 1421.69, mean_step_time 1439.29, it/s 0.69
  6475. [0130 18:02:53 @multigpu.py:323] [p0576]  step: count(1378), step_time 1407.45, mean_step_time 1436.31, it/s 0.7
  6476. [0130 18:02:53 @multigpu.py:323] [p0115]  step: count(1354), step_time 1433.65, mean_step_time 1435.15, it/s 0.7
  6477. [0130 18:02:54 @multigpu.py:323] [p0574]  step: count(1370), step_time 1417.29, mean_step_time 1438.07, it/s 0.7
  6478. [0130 18:02:54 @multigpu.py:323] [p0576]  step: count(1379), step_time 1367.25, mean_step_time 1435.32, it/s 0.7
  6479. [0130 18:02:54 @multigpu.py:323] [p0115]  step: count(1355), step_time 1423.47, mean_step_time 1434.92, it/s 0.7
  6480. [0130 18:02:55 @multigpu.py:323] [p0574]  step: count(1371), step_time 1411.08, mean_step_time 1437.89, it/s 0.7
  6481. [0130 18:02:55 @multigpu.py:323] [p0576]  step: count(1380), step_time 1443.27, mean_step_time 1432.29, it/s 0.7
  6482. [0130 18:02:55 @multigpu.py:323] [p0115]  step: count(1356), step_time 1474.0, mean_step_time 1437.46, it/s 0.7
  6483. [0130 18:02:57 @multigpu.py:323] [p0576]  step: count(1381), step_time 1368.81, mean_step_time 1430.52, it/s 0.7
  6484. [0130 18:02:57 @multigpu.py:323] [p0574]  step: count(1372), step_time 1491.64, mean_step_time 1442.22, it/s 0.69
  6485. [0130 18:02:57 @multigpu.py:323] [p0115]  step: count(1357), step_time 1414.76, mean_step_time 1435.68, it/s 0.7
  6486. [0130 18:02:58 @multigpu.py:323] [p0576]  step: count(1382), step_time 1484.48, mean_step_time 1432.16, it/s 0.7
  6487. [0130 18:02:58 @multigpu.py:323] [p0574]  step: count(1373), step_time 1429.34, mean_step_time 1442.93, it/s 0.69
  6488. [0130 18:02:58 @multigpu.py:323] [p0115]  step: count(1358), step_time 1513.46, mean_step_time 1439.3, it/s 0.69
  6489. [0130 18:03:00 @multigpu.py:323] [p0576]  step: count(1383), step_time 1430.51, mean_step_time 1433.25, it/s 0.7
  6490. [0130 18:03:00 @multigpu.py:323] [p0574]  step: count(1374), step_time 1436.92, mean_step_time 1445.26, it/s 0.69
  6491. [0130 18:03:00 @multigpu.py:323] [p0115]  step: count(1359), step_time 1418.1, mean_step_time 1439.53, it/s 0.69
  6492. [0130 18:03:01 @multigpu.py:323] [p0576]  step: count(1384), step_time 1404.53, mean_step_time 1431.75, it/s 0.7
  6493. [0130 18:03:01 @multigpu.py:323] [p0574]  step: count(1375), step_time 1408.26, mean_step_time 1443.65, it/s 0.69
  6494. [0130 18:03:01 @multigpu.py:323] [p0115]  step: count(1360), step_time 1415.47, mean_step_time 1435.57, it/s 0.7
  6495. [0130 18:03:03 @multigpu.py:323] [p0576]  step: count(1385), step_time 1442.41, mean_step_time 1433.11, it/s 0.7
  6496. [0130 18:03:03 @multigpu.py:323] [p0574]  step: count(1376), step_time 1428.69, mean_step_time 1444.87, it/s 0.69
  6497. [0130 18:03:03 @multigpu.py:323] [p0115]  step: count(1361), step_time 1467.4, mean_step_time 1438.32, it/s 0.7
  6498. sending to address tcp://p0112:61216
  6499. ##### Sending to neptune:  online_score :  0.550689121617 , 1.9 #####
  6500. [u'online', 1.9]
  6501. receiving
  6502. sending to address tcp://p0112:61216
  6503. ##### Sending to neptune:  online_score :  0.550788611637 , 2.2 #####
  6504. [u'online', 2.2]
  6505. receiving
  6506. [0130 18:03:04 @multigpu.py:323] [p0576]  step: count(1386), step_time 1435.36, mean_step_time 1434.46, it/s 0.7
  6507. [0130 18:03:04 @multigpu.py:323] [p0574]  step: count(1377), step_time 1426.97, mean_step_time 1446.86, it/s 0.69
  6508. [0130 18:03:04 @multigpu.py:323] [p0115]  step: count(1362), step_time 1378.13, mean_step_time 1435.59, it/s 0.7
  6509. [0130 18:03:05 @multigpu.py:323] [p0574]  step: count(1378), step_time 1406.52, mean_step_time 1446.39, it/s 0.69
  6510. [0130 18:03:05 @multigpu.py:323] [p0576]  step: count(1387), step_time 1427.56, mean_step_time 1434.18, it/s 0.7
  6511. [0130 18:03:05 @multigpu.py:323] [p0115]  step: count(1363), step_time 1406.02, mean_step_time 1433.46, it/s 0.7
  6512. [0130 18:03:07 @multigpu.py:323] [p0576]  step: count(1388), step_time 1447.5, mean_step_time 1434.72, it/s 0.7
  6513. [0130 18:03:07 @multigpu.py:323] [p0574]  step: count(1379), step_time 1476.22, mean_step_time 1446.0, it/s 0.69
  6514. [0130 18:03:07 @multigpu.py:323] [p0115]  step: count(1364), step_time 1421.08, mean_step_time 1432.1, it/s 0.7
  6515. [0130 18:03:08 @multigpu.py:323] [p0576]  step: count(1389), step_time 1389.55, mean_step_time 1433.97, it/s 0.7
  6516. [0130 18:03:08 @multigpu.py:323] [p0574]  step: count(1380), step_time 1435.9, mean_step_time 1446.45, it/s 0.69
  6517. [0130 18:03:08 @multigpu.py:323] [p0115]  step: count(1365), step_time 1455.47, mean_step_time 1434.19, it/s 0.7
  6518. [0130 18:03:10 @multigpu.py:323] [p0576]  step: count(1390), step_time 1399.11, mean_step_time 1433.06, it/s 0.7
  6519. [0130 18:03:10 @multigpu.py:323] [p0115]  step: count(1366), step_time 1419.1, mean_step_time 1433.25, it/s 0.7
  6520. [0130 18:03:10 @multigpu.py:323] [p0574]  step: count(1381), step_time 1518.95, mean_step_time 1453.77, it/s 0.69
  6521. [0130 18:03:11 @multigpu.py:323] [p0576]  step: count(1391), step_time 1394.68, mean_step_time 1433.95, it/s 0.7
  6522. [0130 18:03:11 @multigpu.py:323] [p0574]  step: count(1382), step_time 1389.37, mean_step_time 1448.9, it/s 0.69
  6523. [0130 18:03:11 @multigpu.py:323] [p0115]  step: count(1367), step_time 1411.2, mean_step_time 1433.48, it/s 0.7
  6524. [0130 18:03:13 @multigpu.py:323] [p0115]  step: count(1368), step_time 1451.68, mean_step_time 1433.96, it/s 0.7
  6525. [0130 18:03:13 @multigpu.py:323] [p0576]  step: count(1392), step_time 1637.72, mean_step_time 1433.93, it/s 0.7
  6526. [0130 18:03:13 @multigpu.py:323] [p0574]  step: count(1383), step_time 1469.12, mean_step_time 1433.83, it/s 0.7
  6527. [0130 18:03:14 @multigpu.py:323] [p0576]  step: count(1393), step_time 1392.96, mean_step_time 1432.13, it/s 0.7
  6528. [0130 18:03:14 @multigpu.py:323] [p0574]  step: count(1384), step_time 1429.08, mean_step_time 1433.79, it/s 0.7
  6529. [0130 18:03:14 @multigpu.py:323] [p0115]  step: count(1369), step_time 1442.2, mean_step_time 1434.45, it/s 0.7
  6530. sending to address tcp://p0112:61216
  6531. ##### Sending to neptune:  online_score :  0.553857938581 , 0.8 #####
  6532. [u'online', 0.8]
  6533. receiving
  6534. [0130 18:03:15 @multigpu.py:323] [p0576]  step: count(1394), step_time 1431.74, mean_step_time 1432.04, it/s 0.7
  6535. [0130 18:03:15 @multigpu.py:323] [p0115]  step: count(1370), step_time 1393.41, mean_step_time 1429.88, it/s 0.7
  6536. [0130 18:03:16 @multigpu.py:323] [p0574]  step: count(1385), step_time 1439.65, mean_step_time 1434.15, it/s 0.7
  6537. [0130 18:03:17 @multigpu.py:323] [p0576]  step: count(1395), step_time 1411.33, mean_step_time 1431.83, it/s 0.7
  6538. [0130 18:03:17 @multigpu.py:323] [p0115]  step: count(1371), step_time 1414.81, mean_step_time 1430.05, it/s 0.7
  6539. [0130 18:03:17 @multigpu.py:323] [p0574]  step: count(1386), step_time 1430.7, mean_step_time 1435.36, it/s 0.7
  6540. [0130 18:03:18 @multigpu.py:323] [p0576]  step: count(1396), step_time 1421.42, mean_step_time 1428.61, it/s 0.7
  6541. [0130 18:03:18 @multigpu.py:323] [p0115]  step: count(1372), step_time 1413.61, mean_step_time 1427.64, it/s 0.7
  6542. [0130 18:03:18 @multigpu.py:323] [p0574]  step: count(1387), step_time 1418.14, mean_step_time 1435.82, it/s 0.7
  6543. [0130 18:03:20 @multigpu.py:323] [p0115]  step: count(1373), step_time 1417.49, mean_step_time 1429.23, it/s 0.7
  6544. [0130 18:03:20 @multigpu.py:323] [p0576]  step: count(1397), step_time 1470.89, mean_step_time 1430.43, it/s 0.7
  6545. [0130 18:03:20 @multigpu.py:323] [p0574]  step: count(1388), step_time 1456.8, mean_step_time 1437.12, it/s 0.7
  6546. [0130 18:03:21 @multigpu.py:323] [p0115]  step: count(1374), step_time 1447.07, mean_step_time 1429.9, it/s 0.7
  6547. [0130 18:03:21 @multigpu.py:323] [p0574]  step: count(1389), step_time 1370.25, mean_step_time 1434.54, it/s 0.7
  6548. [0130 18:03:21 @multigpu.py:323] [p0576]  step: count(1398), step_time 1519.05, mean_step_time 1436.01, it/s 0.7
  6549. [0130 18:03:23 @multigpu.py:323] [p0115]  step: count(1375), step_time 1392.37, mean_step_time 1428.34, it/s 0.7
  6550. [0130 18:03:23 @multigpu.py:323] [p0574]  step: count(1390), step_time 1445.91, mean_step_time 1435.98, it/s 0.7
  6551. [0130 18:03:23 @multigpu.py:323] [p0576]  step: count(1399), step_time 1382.11, mean_step_time 1436.75, it/s 0.7
  6552. [0130 18:03:24 @multigpu.py:323] [p0115]  step: count(1376), step_time 1428.37, mean_step_time 1426.06, it/s 0.7
  6553. [0130 18:03:24 @multigpu.py:323] [p0574]  step: count(1391), step_time 1393.81, mean_step_time 1435.11, it/s 0.7
  6554. [0130 18:03:24 @multigpu.py:323] [p0576]  step: count(1400), step_time 1472.93, mean_step_time 1438.23, it/s 0.7
  6555. sending debugging info...
  6556. sending to address tcp://p0112:61216
  6557. ##### Sending to neptune:  mean_delay :  0.556543689966 , 0.0 #####
  6558. ##### Sending to neptune:  max_delay :  0.556543689966 , -0.0 #####
  6559. sending to address tcp://p0112:61216
  6560. ##### Sending to neptune:  min_delay :  0.556543689966 , -0.0 #####
  6561. [u'delays', [0.0, -0.0, -0.0]]
  6562. receiving
  6563. ##### Sending to neptune:  cost :  0.556544211904 , -0.00979332625866 #####
  6564. ##### Sending to neptune:  policy_loss :  0.556544211904 , -0.407843053341 #####
  6565. sending to address tcp://p0112:61216
  6566. ##### Sending to neptune:  xentropy_loss :  0.556544211904 , -2.28586602211 #####
  6567. ##### Sending to neptune:  value_loss :  0.556544211904 , 1.44016361237 #####
  6568. ##### Sending to neptune:  advantage :  0.556544211904 , 0.00166430650279 #####
  6569. ##### Sending to neptune:  pred_reward :  0.556544211904 , 0.450704455376 #####
  6570. ##### Sending to neptune:  max_logit :  0.556544211904 , 0.210775285959 #####
  6571. [u'loss', -0.009793326258659363, -0.40784305334091187, -2.2858660221099854, 1.4401636123657227, 0.001664306502789259, 0.4507044553756714, 0.21077528595924377]
  6572. receiving
  6573. ##### Sending to neptune:  active_relus :  0.556544706888 , 9774847.18 #####
  6574. ##### Sending to neptune:  dp_per_s :  0.556544706888 , 89.2697802076 #####
  6575. [u'other', 9774847.18, 89.26978020759367]
  6576. receiving
  6577. [0130 18:03:25 @multigpu.py:323] [p0115]  step: count(1377), step_time 1460.58, mean_step_time 1428.35, it/s 0.7
  6578. [0130 18:03:25 @multigpu.py:323] [p0574]  step: count(1392), step_time 1452.28, mean_step_time 1433.14, it/s 0.7
  6579. [0130 18:03:26 @multigpu.py:323] [p0576]  step: count(1401), step_time 1385.76, mean_step_time 1439.08, it/s 0.69
  6580. [0130 18:03:27 @multigpu.py:323] [p0574]  step: count(1393), step_time 1385.9, mean_step_time 1430.97, it/s 0.7
  6581. [0130 18:03:27 @multigpu.py:323] [p0115]  step: count(1378), step_time 1452.95, mean_step_time 1425.33, it/s 0.7
  6582. [0130 18:03:27 @multigpu.py:323] [p0576]  step: count(1402), step_time 1407.95, mean_step_time 1435.25, it/s 0.7
  6583. sending to address tcp://p0112:61216
  6584. ##### Sending to neptune:  online_score :  0.557487060494 , 1.7 #####
  6585. [u'online', 1.7]
  6586. ##### Sending to neptune:  active_workers :  0.557487142483 , 3 #####
  6587. receiving
  6588. [0130 18:03:28 @multigpu.py:323] [p0574]  step: count(1394), step_time 1447.19, mean_step_time 1431.49, it/s 0.7
  6589. [0130 18:03:28 @multigpu.py:323] [p0115]  step: count(1379), step_time 1414.04, mean_step_time 1425.12, it/s 0.7
  6590. [0130 18:03:28 @multigpu.py:323] [p0576]  step: count(1403), step_time 1396.03, mean_step_time 1433.53, it/s 0.7
  6591. [0130 18:03:30 @multigpu.py:323] [p0574]  step: count(1395), step_time 1396.59, mean_step_time 1430.9, it/s 0.7
  6592. [0130 18:03:30 @multigpu.py:323] [p0576]  step: count(1404), step_time 1405.56, mean_step_time 1433.58, it/s 0.7
  6593. [0130 18:03:30 @multigpu.py:323] [p0115]  step: count(1380), step_time 1432.02, mean_step_time 1425.95, it/s 0.7
  6594. [0130 18:03:31 @multigpu.py:323] [p0574]  step: count(1396), step_time 1405.79, mean_step_time 1429.76, it/s 0.7
  6595. [0130 18:03:31 @multigpu.py:323] [p0576]  step: count(1405), step_time 1407.6, mean_step_time 1431.84, it/s 0.7
  6596. [0130 18:03:31 @multigpu.py:323] [p0115]  step: count(1381), step_time 1451.89, mean_step_time 1425.17, it/s 0.7
  6597. [0130 18:03:33 @multigpu.py:323] [p0576]  step: count(1406), step_time 1429.22, mean_step_time 1431.53, it/s 0.7
  6598. [0130 18:03:33 @multigpu.py:323] [p0574]  step: count(1397), step_time 1462.66, mean_step_time 1431.54, it/s 0.7
  6599. [0130 18:03:33 @multigpu.py:323] [p0115]  step: count(1382), step_time 1407.0, mean_step_time 1426.62, it/s 0.7
  6600. [0130 18:03:34 @multigpu.py:323] [p0574]  step: count(1398), step_time 1379.74, mean_step_time 1430.2, it/s 0.7
  6601. [0130 18:03:34 @multigpu.py:323] [p0115]  step: count(1383), step_time 1406.98, mean_step_time 1426.67, it/s 0.7
  6602. [0130 18:03:34 @multigpu.py:323] [p0576]  step: count(1407), step_time 1441.96, mean_step_time 1432.25, it/s 0.7
  6603. [0130 18:03:35 @multigpu.py:323] [p0574]  step: count(1399), step_time 1403.81, mean_step_time 1426.58, it/s 0.7
  6604. [0130 18:03:35 @multigpu.py:323] [p0576]  step: count(1408), step_time 1380.08, mean_step_time 1428.88, it/s 0.7
  6605. [0130 18:03:35 @multigpu.py:323] [p0115]  step: count(1384), step_time 1433.42, mean_step_time 1427.28, it/s 0.7
  6606. [0130 18:03:37 @multigpu.py:323] [p0576]  step: count(1409), step_time 1406.49, mean_step_time 1429.73, it/s 0.7
  6607. [0130 18:03:37 @multigpu.py:323] [p0574]  step: count(1400), step_time 1479.25, mean_step_time 1428.75, it/s 0.7
  6608. sending debugging info...
  6609. sending to address tcp://p0112:61216
  6610. ##### Sending to neptune:  mean_delay :  0.560071513851 , 0.0 #####
  6611. sending to address tcp://p0112:61216
  6612. ##### Sending to neptune:  max_delay :  0.560071513851 , -0.0 #####
  6613. ##### Sending to neptune:  min_delay :  0.560071513851 , -0.0 #####
  6614. [u'delays', [0.0, -0.0, -0.0]]
  6615. receiving
  6616. ##### Sending to neptune:  cost :  0.560071992477 , -0.00401578377932 #####
  6617. sending to address tcp://p0112:61216
  6618. ##### Sending to neptune:  policy_loss :  0.560071992477 , 0.167572632432 #####
  6619. ##### Sending to neptune:  xentropy_loss :  0.560071992477 , -2.28561639786 #####
  6620. ##### Sending to neptune:  value_loss :  0.560071992477 , 1.60402357578 #####
  6621. ##### Sending to neptune:  advantage :  0.560071992477 , -0.000748563557863 #####
  6622. ##### Sending to neptune:  pred_reward :  0.560071992477 , 0.458303689957 #####
  6623. ##### Sending to neptune:  max_logit :  0.560071992477 , 0.2134732306 #####
  6624. [u'loss', -0.004015783779323101, 0.16757263243198395, -2.285616397857666, 1.6040235757827759, -0.0007485635578632355, 0.45830368995666504, 0.21347323060035706]
  6625. receiving
  6626. ##### Sending to neptune:  active_relus :  0.560072435273 , 9771696.53 #####
  6627. ##### Sending to neptune:  dp_per_s :  0.560072435273 , 89.2691596223 #####
  6628. [u'other', 9771696.53, 89.26915962233971]
  6629. receiving
  6630. [0130 18:03:37 @multigpu.py:323] [p0115]  step: count(1385), step_time 1439.52, mean_step_time 1426.49, it/s 0.7
  6631. sending to address tcp://p0112:61216
  6632. ##### Sending to neptune:  online_score :  0.560282720261 , 2.1 #####
  6633. [u'online', 2.1]
  6634. receiving
  6635. [0130 18:03:38 @multigpu.py:323] [p0576]  step: count(1410), step_time 1412.38, mean_step_time 1430.39, it/s 0.7
  6636. [0130 18:03:38 @multigpu.py:323] [p0574]  step: count(1401), step_time 1418.35, mean_step_time 1423.72, it/s 0.7
  6637. [0130 18:03:38 @multigpu.py:323] [p0115]  step: count(1386), step_time 1427.55, mean_step_time 1426.91, it/s 0.7
  6638. [0130 18:03:40 @multigpu.py:323] [p0576]  step: count(1411), step_time 1410.81, mean_step_time 1431.2, it/s 0.7
  6639. [0130 18:03:40 @multigpu.py:323] [p0574]  step: count(1402), step_time 1430.52, mean_step_time 1425.78, it/s 0.7
  6640. [0130 18:03:40 @multigpu.py:323] [p0115]  step: count(1387), step_time 1387.21, mean_step_time 1425.71, it/s 0.7
  6641. [0130 18:03:41 @multigpu.py:323] [p0115]  step: count(1388), step_time 1451.92, mean_step_time 1425.72, it/s 0.7
  6642. [0130 18:03:41 @multigpu.py:323] [p0574]  step: count(1403), step_time 1468.66, mean_step_time 1425.75, it/s 0.7
  6643. [0130 18:03:41 @multigpu.py:323] [p0576]  step: count(1412), step_time 1529.78, mean_step_time 1425.8, it/s 0.7
  6644. [0130 18:03:43 @multigpu.py:323] [p0574]  step: count(1404), step_time 1376.73, mean_step_time 1423.14, it/s 0.7
  6645. [0130 18:03:43 @multigpu.py:323] [p0576]  step: count(1413), step_time 1411.29, mean_step_time 1426.72, it/s 0.7
  6646. [0130 18:03:43 @multigpu.py:323] [p0115]  step: count(1389), step_time 1436.06, mean_step_time 1425.41, it/s 0.7
  6647. [0130 18:03:44 @multigpu.py:323] [p0574]  step: count(1405), step_time 1419.54, mean_step_time 1422.13, it/s 0.7
  6648. [0130 18:03:44 @multigpu.py:323] [p0115]  step: count(1390), step_time 1406.9, mean_step_time 1426.09, it/s 0.7
  6649. [0130 18:03:44 @multigpu.py:323] [p0576]  step: count(1414), step_time 1453.25, mean_step_time 1427.79, it/s 0.7
  6650. [0130 18:03:45 @multigpu.py:323] [p0574]  step: count(1406), step_time 1442.56, mean_step_time 1422.72, it/s 0.7
  6651. [0130 18:03:45 @multigpu.py:323] [p0115]  step: count(1391), step_time 1404.4, mean_step_time 1425.57, it/s 0.7
  6652. [0130 18:03:45 @multigpu.py:323] [p0576]  step: count(1415), step_time 1411.3, mean_step_time 1427.79, it/s 0.7
  6653. [0130 18:03:47 @multigpu.py:323] [p0574]  step: count(1407), step_time 1372.41, mean_step_time 1420.44, it/s 0.7
  6654. [0130 18:03:47 @multigpu.py:323] [p0115]  step: count(1392), step_time 1419.39, mean_step_time 1425.86, it/s 0.7
  6655. [0130 18:03:47 @multigpu.py:323] [p0576]  step: count(1416), step_time 1451.4, mean_step_time 1429.29, it/s 0.7
  6656. [0130 18:03:48 @multigpu.py:323] [p0574]  step: count(1408), step_time 1446.35, mean_step_time 1419.91, it/s 0.7
  6657. [0130 18:03:48 @multigpu.py:323] [p0115]  step: count(1393), step_time 1410.03, mean_step_time 1425.48, it/s 0.7
  6658. [0130 18:03:48 @multigpu.py:323] [p0576]  step: count(1417), step_time 1358.56, mean_step_time 1423.68, it/s 0.7
  6659. [0130 18:03:50 @multigpu.py:323] [p0576]  step: count(1418), step_time 1427.34, mean_step_time 1419.09, it/s 0.7
  6660. [0130 18:03:50 @multigpu.py:323] [p0115]  step: count(1394), step_time 1481.06, mean_step_time 1427.18, it/s 0.7
  6661. [0130 18:03:50 @multigpu.py:323] [p0574]  step: count(1409), step_time 1503.11, mean_step_time 1426.56, it/s 0.7
  6662. [0130 18:03:51 @multigpu.py:323] [p0115]  step: count(1395), step_time 1411.36, mean_step_time 1428.13, it/s 0.7
  6663. [0130 18:03:51 @multigpu.py:323] [p0576]  step: count(1419), step_time 1463.2, mean_step_time 1423.14, it/s 0.7
  6664. sending to address tcp://p0112:61216
  6665. ##### Sending to neptune:  online_score :  0.564045781096 , 2.1 #####
  6666. [u'online', 2.1]
  6667. receiving
  6668. [0130 18:03:51 @multigpu.py:323] [p0574]  step: count(1410), step_time 1455.92, mean_step_time 1427.06, it/s 0.7
  6669. [0130 18:03:53 @multigpu.py:323] [p0576]  step: count(1420), step_time 1406.7, mean_step_time 1419.83, it/s 0.7
  6670. [0130 18:03:53 @multigpu.py:323] [p0115]  step: count(1396), step_time 1418.38, mean_step_time 1427.63, it/s 0.7
  6671. [0130 18:03:53 @multigpu.py:323] [p0574]  step: count(1411), step_time 1499.37, mean_step_time 1432.34, it/s 0.7
  6672. [0130 18:03:54 @multigpu.py:323] [p0115]  step: count(1397), step_time 1440.01, mean_step_time 1426.6, it/s 0.7
  6673. [0130 18:03:54 @multigpu.py:323] [p0576]  step: count(1421), step_time 1460.03, mean_step_time 1423.55, it/s 0.7
  6674.  42%|####2     |421/1000[1[0130 18:03:54 @multigpu.py:323] [p0574]  step: count(1412), step_time 1433.92, mean_step_time 1431.42, it/s 0.7
  6675. [0130 18:03:55 @multigpu.py:323] [p0576]  step: count(1422), step_time 1431.75, mean_step_time 1424.74, it/s 0.7
  6676. [0130 18:03:56 @multigpu.py:323] [p0115]  step: count(1398), step_time 1491.19, mean_step_time 1428.52, it/s 0.7
  6677. [0130 18:03:56 @multigpu.py:323] [p0574]  step: count(1413), step_time 1429.86, mean_step_time 1433.62, it/s 0.7
  6678. [0130 18:03:57 @multigpu.py:323] [p0576]  step: count(1423), step_time 1400.64, mean_step_time 1424.97, it/s 0.7
  6679. [0130 18:03:57 @multigpu.py:323] [p0115]  step: count(1399), step_time 1451.76, mean_step_time 1430.4, it/s 0.7
  6680. [0130 18:03:57 @multigpu.py:323] [p0574]  step: count(1414), step_time 1432.31, mean_step_time 1432.87, it/s 0.7
  6681. [0130 18:03:58 @multigpu.py:323] [p0576]  step: count(1424), step_time 1441.02, mean_step_time 1426.74, it/s 0.7
  6682. [0130 18:03:58 @multigpu.py:323] [p0115]  step: count(1400), step_time 1432.93, mean_step_time 1430.45, it/s 0.7
  6683. sending debugging info...
  6684. sending to address tcp://p0112:61216
  6685. ##### Sending to neptune:  mean_delay :  0.566052081651 , 0.0 #####
  6686. sending to address tcp://p0112:61216
  6687. ##### Sending to neptune:  max_delay :  0.566052081651 , -0.0 #####
  6688. ##### Sending to neptune:  min_delay :  0.566052081651 , -0.0 #####
  6689. [u'delays', [0.0, -0.0, -0.0]]
  6690. receiving
  6691. ##### Sending to neptune:  cost :  0.566052584383 , -0.00693348702043 #####
  6692. sending to address tcp://p0112:61216
  6693. ##### Sending to neptune:  policy_loss :  0.566052584383 , 0.0658853203058 #####
  6694. ##### Sending to neptune:  xentropy_loss :  0.566052584383 , -2.28544616699 #####
  6695. ##### Sending to neptune:  value_loss :  0.566052584383 , 1.33207464218 #####
  6696. ##### Sending to neptune:  advantage :  0.566052584383 , -0.000360454781912 #####
  6697. ##### Sending to neptune:  pred_reward :  0.566052584383 , 0.440528899431 #####
  6698. ##### Sending to neptune:  max_logit :  0.566052584383 , 0.215451806784 #####
  6699. [u'loss', -0.006933487020432949, 0.06588532030582428, -2.2854461669921875, 1.3320746421813965, -0.0003604547819122672, 0.44052889943122864, 0.21545180678367615]
  6700. receiving
  6701. ##### Sending to neptune:  active_relus :  0.566053047445 , 9768145.36 #####
  6702. ##### Sending to neptune:  dp_per_s :  0.566053047445 , 89.2624328751 #####
  6703. [u'other', 9768145.36, 89.26243287508888]
  6704. receiving
  6705. [0130 18:03:58 @multigpu.py:323] [p0574]  step: count(1415), step_time 1433.64, mean_step_time 1434.72, it/s 0.7
  6706. sending to address tcp://p0112:61216
  6707. ##### Sending to neptune:  online_score :  0.566195656061 , 1.7 #####
  6708. [u'online', 1.7]
  6709. receiving
  6710. [0130 18:04:00 @multigpu.py:323] [p0576]  step: count(1425), step_time 1418.71, mean_step_time 1427.29, it/s 0.7
  6711. [0130 18:04:00 @multigpu.py:323] [p0115]  step: count(1401), step_time 1408.59, mean_step_time 1428.28, it/s 0.7
  6712. [0130 18:04:00 @multigpu.py:323] [p0574]  step: count(1416), step_time 1428.78, mean_step_time 1435.87, it/s 0.7
  6713. [0130 18:04:01 @multigpu.py:323] [p0576]  step: count(1426), step_time 1384.2, mean_step_time 1425.04, it/s 0.7
  6714. [0130 18:04:01 @multigpu.py:323] [p0115]  step: count(1402), step_time 1472.46, mean_step_time 1431.56, it/s 0.7
  6715. [0130 18:04:01 @multigpu.py:323] [p0574]  step: count(1417), step_time 1427.13, mean_step_time 1434.1, it/s 0.7
  6716. [0130 18:04:03 @multigpu.py:323] [p0576]  step: count(1427), step_time 1432.02, mean_step_time 1424.55, it/s 0.7
  6717. [0130 18:04:03 @multigpu.py:323] [p0115]  step: count(1403), step_time 1397.03, mean_step_time 1431.06, it/s 0.7
  6718. [0130 18:04:03 @multigpu.py:323] [p0574]  step: count(1418), step_time 1401.49, mean_step_time 1435.18, it/s 0.7
  6719. sending to address tcp://p0112:61216
  6720. ##### Sending to neptune:  online_score :  0.567442069716 , 2.2 #####
  6721. [u'online', 2.2]
  6722. receiving
  6723. [0130 18:04:04 @multigpu.py:323] [p0576]  step: count(1428), step_time 1388.66, mean_step_time 1424.98, it/s 0.7
  6724. [0130 18:04:04 @multigpu.py:323] [p0115]  step: count(1404), step_time 1416.48, mean_step_time 1430.21, it/s 0.7
  6725. [0130 18:04:04 @multigpu.py:323] [p0574]  step: count(1419), step_time 1474.8, mean_step_time 1438.73, it/s 0.7
  6726. [0130 18:04:05 @multigpu.py:323] [p0576]  step: count(1429), step_time 1433.08, mean_step_time 1426.31, it/s 0.7
  6727. [0130 18:04:06 @multigpu.py:323] [p0115]  step: count(1405), step_time 1485.3, mean_step_time 1432.5, it/s 0.7
  6728. [0130 18:04:06 @multigpu.py:323] [p0574]  step: count(1420), step_time 1429.36, mean_step_time 1436.24, it/s 0.7
  6729. [0130 18:04:07 @multigpu.py:323] [p0576]  step: count(1430), step_time 1402.01, mean_step_time 1425.79, it/s 0.7
  6730. [0130 18:04:07 @multigpu.py:323] [p0115]  step: count(1406), step_time 1430.27, mean_step_time 1432.64, it/s 0.7
  6731. [0130 18:04:07 @multigpu.py:323] [p0574]  step: count(1421), step_time 1446.44, mean_step_time 1437.64, it/s 0.7
  6732.  42%|####2     |421/1000[1[0130 18:04:08 @multigpu.py:323] [p0576]  step: count(1431), step_time 1437.82, mean_step_time 1427.14, it/s 0.7
  6733. [0130 18:04:08 @multigpu.py:323] [p0574]  step: count(1422), step_time 1387.96, mean_step_time 1435.52, it/s 0.7
  6734. [0130 18:04:08 @multigpu.py:323] [p0115]  step: count(1407), step_time 1463.99, mean_step_time 1436.48, it/s 0.7
  6735. [0130 18:04:10 @multigpu.py:323] [p0115]  step: count(1408), step_time 1417.2, mean_step_time 1434.74, it/s 0.7
  6736. [0130 18:04:10 @multigpu.py:323] [p0576]  step: count(1432), step_time 1683.99, mean_step_time 1434.85, it/s 0.7
  6737. [0130 18:04:10 @multigpu.py:323] [p0574]  step: count(1423), step_time 1453.76, mean_step_time 1434.77, it/s 0.7
  6738. [0130 18:04:11 @multigpu.py:323] [p0574]  step: count(1424), step_time 1397.23, mean_step_time 1435.8, it/s 0.7
  6739. [0130 18:04:11 @multigpu.py:323] [p0576]  step: count(1433), step_time 1445.1, mean_step_time 1436.54, it/s 0.7
  6740. [0130 18:04:11 @multigpu.py:323] [p0115]  step: count(1409), step_time 1479.96, mean_step_time 1436.93, it/s 0.7
  6741. [0130 18:04:13 @multigpu.py:323] [p0574]  step: count(1425), step_time 1455.64, mean_step_time 1437.6, it/s 0.7
  6742. [0130 18:04:13 @multigpu.py:323] [p0576]  step: count(1434), step_time 1429.56, mean_step_time 1435.35, it/s 0.7
  6743. [0130 18:04:13 @multigpu.py:323] [p0115]  step: count(1410), step_time 1414.51, mean_step_time 1437.31, it/s 0.7
  6744. [0130 18:04:14 @multigpu.py:323] [p0574]  step: count(1426), step_time 1404.97, mean_step_time 1435.72, it/s 0.7
  6745. [0130 18:04:14 @multigpu.py:323] [p0576]  step: count(1435), step_time 1443.08, mean_step_time 1436.94, it/s 0.7
  6746. [0130 18:04:14 @multigpu.py:323] [p0115]  step: count(1411), step_time 1460.86, mean_step_time 1440.14, it/s 0.69
  6747. [0130 18:04:16 @multigpu.py:323] [p0574]  step: count(1427), step_time 1421.89, mean_step_time 1438.2, it/s 0.7
  6748. [0130 18:04:16 @multigpu.py:323] [p0115]  step: count(1412), step_time 1443.55, mean_step_time 1441.35, it/s 0.69
  6749. [0130 18:04:16 @multigpu.py:323] [p0576]  step: count(1436), step_time 1507.11, mean_step_time 1439.73, it/s 0.69
  6750. [0130 18:04:17 @multigpu.py:323] [p0574]  step: count(1428), step_time 1436.85, mean_step_time 1437.72, it/s 0.7
  6751. [0130 18:04:17 @multigpu.py:323] [p0115]  step: count(1413), step_time 1404.07, mean_step_time 1441.05, it/s 0.69
  6752. [0130 18:04:17 @multigpu.py:323] [p0576]  step: count(1437), step_time 1402.3, mean_step_time 1441.92, it/s 0.69
  6753. [0130 18:04:18 @multigpu.py:323] [p0574]  step: count(1429), step_time 1403.43, mean_step_time 1432.74, it/s 0.7
  6754. [0130 18:04:19 @multigpu.py:323] [p0115]  step: count(1414), step_time 1450.1, mean_step_time 1439.5, it/s 0.69
  6755. [0130 18:04:19 @multigpu.py:323] [p0576]  step: count(1438), step_time 1429.5, mean_step_time 1442.02, it/s 0.69
  6756. [0130 18:04:20 @multigpu.py:323] [p0574]  step: count(1430), step_time 1454.1, mean_step_time 1432.65, it/s 0.7
  6757. [0130 18:04:20 @multigpu.py:323] [p0576]  step: count(1439), step_time 1378.6, mean_step_time 1437.79, it/s 0.7
  6758. [0130 18:04:20 @multigpu.py:323] [p0115]  step: count(1415), step_time 1405.54, mean_step_time 1439.21, it/s 0.69
  6759. [0130 18:04:21 @multigpu.py:323] [p0574]  step: count(1431), step_time 1439.17, mean_step_time 1429.64, it/s 0.7
  6760. [0130 18:04:21 @multigpu.py:323] [p0115]  step: count(1416), step_time 1441.28, mean_step_time 1440.35, it/s 0.69
  6761. [0130 18:04:21 @multigpu.py:323] [p0576]  step: count(1440), step_time 1499.94, mean_step_time 1442.46, it/s 0.69
  6762. [0130 18:04:23 @multigpu.py:323] [p0574]  step: count(1432), step_time 1418.43, mean_step_time 1428.86, it/s 0.7
  6763. [0130 18:04:23 @multigpu.py:323] [p0115]  step: count(1417), step_time 1441.13, mean_step_time 1440.41, it/s 0.69
  6764. [0130 18:04:23 @multigpu.py:323] [p0576]  step: count(1441), step_time 1418.78, mean_step_time 1440.39, it/s 0.69
  6765. sending to address tcp://p0112:61216
  6766. ##### Sending to neptune:  online_score :  0.57297924108 , 2.1 #####
  6767. [u'online', 2.1]
  6768. receiving
  6769. [0130 18:04:24 @multigpu.py:323] [p0574]  step: count(1433), step_time 1421.44, mean_step_time 1428.44, it/s 0.7
  6770. [0130 18:04:24 @multigpu.py:323] [p0115]  step: count(1418), step_time 1364.18, mean_step_time 1434.06, it/s 0.7
  6771. [0130 18:04:24 @multigpu.py:323] [p0576]  step: count(1442), step_time 1402.34, mean_step_time 1438.92, it/s 0.69
  6772. [0130 18:04:26 @multigpu.py:323] [p0574]  step: count(1434), step_time 1469.75, mean_step_time 1430.31, it/s 0.7
  6773. [0130 18:04:26 @multigpu.py:323] [p0115]  step: count(1419), step_time 1464.25, mean_step_time 1434.68, it/s 0.7
  6774. [0130 18:04:26 @multigpu.py:323] [p0576]  step: count(1443), step_time 1439.18, mean_step_time 1440.85, it/s 0.69
  6775. [0130 18:04:27 @multigpu.py:323] [p0574]  step: count(1435), step_time 1390.89, mean_step_time 1428.18, it/s 0.7
  6776. [0130 18:04:27 @multigpu.py:323] [p0576]  step: count(1444), step_time 1393.86, mean_step_time 1438.49, it/s 0.7
  6777. [0130 18:04:27 @multigpu.py:323] [p0115]  step: count(1420), step_time 1461.02, mean_step_time 1436.09, it/s 0.7
  6778.  42%|####2     |420/1000[1[0130 18:04:28 @multigpu.py:323] [p0574]  step: count(1436), step_time 1398.75, mean_step_time 1426.67, it/s 0.7
  6779. [0130 18:04:29 @multigpu.py:323] [p0576]  step: count(1445), step_time 1473.53, mean_step_time 1441.23, it/s 0.69
  6780. [0130 18:04:29 @multigpu.py:323] [p0115]  step: count(1421), step_time 1498.68, mean_step_time 1440.59, it/s 0.69
  6781. sending to address tcp://p0112:61216
  6782. ##### Sending to neptune:  online_score :  0.574587674141 , 1.3 #####
  6783. [u'online', 1.3]
  6784. ##### Sending to neptune:  active_workers :  0.57458770944 , 3 #####
  6785. receiving
  6786. [0130 18:04:30 @multigpu.py:323] [p0574]  step: count(1437), step_time 1419.64, mean_step_time 1426.3, it/s 0.7
  6787. [0130 18:04:30 @multigpu.py:323] [p0115]  step: count(1422), step_time 1420.56, mean_step_time 1438.0, it/s 0.7
  6788. [0130 18:04:30 @multigpu.py:323] [p0576]  step: count(1446), step_time 1496.92, mean_step_time 1446.87, it/s 0.69
  6789. sending to address tcp://p0112:61216
  6790. ##### Sending to neptune:  online_score :  0.574955003858 , 1.9 #####
  6791. [u'online', 1.9]
  6792. receiving
  6793. [0130 18:04:31 @multigpu.py:323] [p0574]  step: count(1438), step_time 1439.34, mean_step_time 1428.19, it/s 0.7
  6794. [0130 18:04:31 @multigpu.py:323] [p0115]  step: count(1423), step_time 1413.13, mean_step_time 1438.8, it/s 0.7
  6795. [0130 18:04:31 @multigpu.py:323] [p0576]  step: count(1447), step_time 1426.12, mean_step_time 1446.57, it/s 0.69
  6796. [0130 18:04:33 @multigpu.py:323] [p0574]  step: count(1439), step_time 1413.37, mean_step_time 1425.12, it/s 0.7
  6797. [0130 18:04:33 @multigpu.py:323] [p0115]  step: count(1424), step_time 1406.32, mean_step_time 1438.29, it/s 0.7
  6798. [0130 18:04:33 @multigpu.py:323] [p0576]  step: count(1448), step_time 1443.4, mean_step_time 1449.31, it/s 0.69
  6799. [0130 18:04:34 @multigpu.py:323] [p0574]  step: count(1440), step_time 1384.12, mean_step_time 1422.86, it/s 0.7
  6800. [0130 18:04:34 @multigpu.py:323] [p0115]  step: count(1425), step_time 1440.45, mean_step_time 1436.05, it/s 0.7
  6801. [0130 18:04:34 @multigpu.py:323] [p0576]  step: count(1449), step_time 1399.53, mean_step_time 1447.63, it/s 0.69
  6802. [0130 18:04:35 @multigpu.py:323] [p0574]  step: count(1441), step_time 1419.58, mean_step_time 1421.51, it/s 0.7
  6803. [0130 18:04:36 @multigpu.py:323] [p0115]  step: count(1426), step_time 1411.7, mean_step_time 1435.12, it/s 0.7
  6804. [0130 18:04:36 @multigpu.py:323] [p0576]  step: count(1450), step_time 1423.99, mean_step_time 1448.73, it/s 0.69
  6805. [0130 18:04:37 @multigpu.py:323] [p0574]  step: count(1442), step_time 1436.64, mean_step_time 1423.95, it/s 0.7
  6806. [0130 18:04:37 @multigpu.py:323] [p0576]  step: count(1451), step_time 1416.57, mean_step_time 1447.67, it/s 0.69
  6807. [0130 18:04:37 @multigpu.py:323] [p0115]  step: count(1427), step_time 1510.37, mean_step_time 1437.44, it/s 0.7
  6808. [0130 18:04:39 @multigpu.py:323] [p0115]  step: count(1428), step_time 1389.38, mean_step_time 1436.05, it/s 0.7
  6809. [0130 18:04:39 @multigpu.py:323] [p0574]  step: count(1443), step_time 1693.65, mean_step_time 1435.94, it/s 0.7
  6810. [0130 18:04:39 @multigpu.py:323] [p0576]  step: count(1452), step_time 1449.86, mean_step_time 1435.96, it/s 0.7
  6811. [0130 18:04:40 @multigpu.py:323] [p0574]  step: count(1444), step_time 1416.81, mean_step_time 1436.92, it/s 0.7
  6812. [0130 18:04:40 @multigpu.py:323] [p0115]  step: count(1429), step_time 1443.53, mean_step_time 1434.23, it/s 0.7
  6813. [0130 18:04:40 @multigpu.py:323] [p0576]  step: count(1453), step_time 1449.07, mean_step_time 1436.16, it/s 0.7
  6814. [0130 18:04:41 @multigpu.py:323] [p0574]  step: count(1445), step_time 1428.49, mean_step_time 1435.56, it/s 0.7
  6815. [0130 18:04:42 @multigpu.py:323] [p0576]  step: count(1454), step_time 1435.61, mean_step_time 1436.46, it/s 0.7
  6816. [0130 18:04:42 @multigpu.py:323] [p0115]  step: count(1430), step_time 1450.77, mean_step_time 1436.04, it/s 0.7
  6817. [0130 18:04:43 @multigpu.py:323] [p0574]  step: count(1446), step_time 1397.11, mean_step_time 1435.17, it/s 0.7
  6818. [0130 18:04:43 @multigpu.py:323] [p0115]  step: count(1431), step_time 1429.27, mean_step_time 1434.46, it/s 0.7
  6819. [0130 18:04:43 @multigpu.py:323] [p0576]  step: count(1455), step_time 1445.67, mean_step_time 1436.59, it/s 0.7
  6820. [0130 18:04:44 @multigpu.py:323] [p0574]  step: count(1447), step_time 1409.48, mean_step_time 1434.55, it/s 0.7
  6821. [0130 18:04:44 @multigpu.py:323] [p0576]  step: count(1456), step_time 1421.41, mean_step_time 1432.31, it/s 0.7
  6822. [0130 18:04:44 @multigpu.py:323] [p0115]  step: count(1432), step_time 1447.53, mean_step_time 1434.66, it/s 0.7
  6823. [0130 18:04:46 @multigpu.py:323] [p0574]  step: count(1448), step_time 1405.16, mean_step_time 1432.97, it/s 0.7
  6824. [0130 18:04:46 @multigpu.py:323] [p0576]  step: count(1457), step_time 1417.23, mean_step_time 1433.06, it/s 0.7
  6825. [0130 18:04:46 @multigpu.py:323] [p0115]  step: count(1433), step_time 1425.03, mean_step_time 1435.71, it/s 0.7
  6826. [0130 18:04:47 @multigpu.py:323] [p0574]  step: count(1449), step_time 1415.3, mean_step_time 1433.56, it/s 0.7
  6827. [0130 18:04:47 @multigpu.py:323] [p0115]  step: count(1434), step_time 1399.97, mean_step_time 1433.2, it/s 0.7
  6828. [0130 18:04:47 @multigpu.py:323] [p0576]  step: count(1458), step_time 1449.31, mean_step_time 1434.05, it/s 0.7
  6829. [0130 18:04:49 @multigpu.py:323] [p0574]  step: count(1450), step_time 1483.76, mean_step_time 1435.04, it/s 0.7
  6830. [0130 18:04:49 @multigpu.py:323] [p0115]  step: count(1435), step_time 1417.79, mean_step_time 1433.82, it/s 0.7
  6831. [0130 18:04:49 @multigpu.py:323] [p0576]  step: count(1459), step_time 1436.14, mean_step_time 1436.92, it/s 0.7
  6832. [0130 18:04:50 @multigpu.py:323] [p0574]  step: count(1451), step_time 1404.69, mean_step_time 1433.32, it/s 0.7
  6833. [0130 18:04:50 @multigpu.py:323] [p0115]  step: count(1436), step_time 1410.73, mean_step_time 1432.29, it/s 0.7
  6834. [0130 18:04:50 @multigpu.py:323] [p0576]  step: count(1460), step_time 1416.0, mean_step_time 1432.73, it/s 0.7
  6835. [0130 18:04:51 @multigpu.py:323] [p0574]  step: count(1452), step_time 1442.82, mean_step_time 1434.54, it/s 0.7
  6836. [0130 18:04:51 @multigpu.py:323] [p0576]  step: count(1461), step_time 1396.68, mean_step_time 1431.62, it/s 0.7
  6837. [0130 18:04:52 @multigpu.py:323] [p0115]  step: count(1437), step_time 1458.91, mean_step_time 1433.18, it/s 0.7
  6838. [0130 18:04:53 @multigpu.py:323] [p0576]  step: count(1462), step_time 1413.84, mean_step_time 1432.2, it/s 0.7
  6839. [0130 18:04:53 @multigpu.py:323] [p0115]  step: count(1438), step_time 1434.71, mean_step_time 1436.7, it/s 0.7
  6840. [0130 18:04:53 @multigpu.py:323] [p0574]  step: count(1453), step_time 1518.88, mean_step_time 1439.41, it/s 0.69
  6841. sending to address tcp://p0112:61216
  6842. ##### Sending to neptune:  online_score :  0.581396578881 , 2.4 #####
  6843. [u'online', 2.4]
  6844. receiving
  6845. [0130 18:04:54 @multigpu.py:323] [p0576]  step: count(1463), step_time 1382.78, mean_step_time 1429.38, it/s 0.7
  6846. [0130 18:04:54 @multigpu.py:323] [p0574]  step: count(1454), step_time 1381.92, mean_step_time 1435.02, it/s 0.7
  6847. [0130 18:04:54 @multigpu.py:323] [p0115]  step: count(1439), step_time 1474.45, mean_step_time 1437.22, it/s 0.7
  6848. sending to address tcp://p0112:61216
  6849. ##### Sending to neptune:  online_score :  0.58178515527 , 1.7 #####
  6850. [u'online', 1.7]
  6851. receiving
  6852. [0130 18:04:56 @multigpu.py:323] [p0576]  step: count(1464), step_time 1446.41, mean_step_time 1432.0, it/s 0.7
  6853. [0130 18:04:56 @multigpu.py:323] [p0574]  step: count(1455), step_time 1437.95, mean_step_time 1437.37, it/s 0.7
  6854. [0130 18:04:56 @multigpu.py:323] [p0115]  step: count(1440), step_time 1431.34, mean_step_time 1435.73, it/s 0.7
  6855. [0130 18:04:57 @multigpu.py:323] [p0576]  step: count(1465), step_time 1396.41, mean_step_time 1428.15, it/s 0.7
  6856. [0130 18:04:57 @multigpu.py:323] [p0574]  step: count(1456), step_time 1366.36, mean_step_time 1435.75, it/s 0.7
  6857. [0130 18:04:57 @multigpu.py:323] [p0115]  step: count(1441), step_time 1456.16, mean_step_time 1433.61, it/s 0.7
  6858. [0130 18:04:59 @multigpu.py:323] [p0576]  step: count(1466), step_time 1406.44, mean_step_time 1423.62, it/s 0.7
  6859. [0130 18:04:59 @multigpu.py:323] [p0574]  step: count(1457), step_time 1428.66, mean_step_time 1436.21, it/s 0.7
  6860. [0130 18:04:59 @multigpu.py:323] [p0115]  step: count(1442), step_time 1452.52, mean_step_time 1435.2, it/s 0.7
  6861. [0130 18:05:00 @multigpu.py:323] [p0574]  step: count(1458), step_time 1429.06, mean_step_time 1435.69, it/s 0.7
  6862. sending to address tcp://p0112:61216
  6863. ##### Sending to neptune:  online_score :  0.583168856634 , 1.6 #####
  6864. [u'online', 1.6]
  6865. receiving
  6866. [0130 18:05:00 @multigpu.py:323] [p0576]  step: count(1467), step_time 1486.09, mean_step_time 1426.62, it/s 0.7
  6867. [0130 18:05:00 @multigpu.py:323] [p0115]  step: count(1443), step_time 1417.34, mean_step_time 1435.41, it/s 0.7
  6868. [0130 18:05:01 @multigpu.py:323] [p0574]  step: count(1459), step_time 1460.65, mean_step_time 1438.06, it/s 0.7
  6869. [0130 18:05:01 @multigpu.py:323] [p0576]  step: count(1468), step_time 1442.72, mean_step_time 1426.59, it/s 0.7
  6870. [0130 18:05:02 @multigpu.py:323] [p0115]  step: count(1444), step_time 1443.77, mean_step_time 1437.29, it/s 0.7
  6871. [0130 18:05:03 @multigpu.py:323] [p0574]  step: count(1460), step_time 1357.34, mean_step_time 1436.72, it/s 0.7
  6872. [0130 18:05:03 @multigpu.py:323] [p0576]  step: count(1469), step_time 1421.67, mean_step_time 1427.69, it/s 0.7
  6873. [0130 18:05:03 @multigpu.py:323] [p0115]  step: count(1445), step_time 1390.52, mean_step_time 1434.79, it/s 0.7
  6874. [0130 18:05:04 @multigpu.py:323] [p0574]  step: count(1461), step_time 1439.14, mean_step_time 1437.69, it/s 0.7
  6875. [0130 18:05:04 @multigpu.py:323] [p0576]  step: count(1470), step_time 1397.91, mean_step_time 1426.39, it/s 0.7
  6876. [0130 18:05:04 @multigpu.py:323] [p0115]  step: count(1446), step_time 1459.09, mean_step_time 1437.16, it/s 0.7
  6877. [0130 18:05:06 @multigpu.py:323] [p0574]  step: count(1462), step_time 1395.75, mean_step_time 1435.65, it/s 0.7
  6878. [0130 18:05:06 @multigpu.py:323] [p0576]  step: count(1471), step_time 1491.13, mean_step_time 1430.12, it/s 0.7
  6879. [0130 18:05:06 @multigpu.py:323] [p0115]  step: count(1447), step_time 1441.38, mean_step_time 1433.71, it/s 0.7
  6880. [0130 18:05:07 @multigpu.py:323] [p0115]  step: count(1448), step_time 1416.11, mean_step_time 1435.05, it/s 0.7
  6881. [0130 18:05:07 @multigpu.py:323] [p0576]  step: count(1472), step_time 1545.41, mean_step_time 1434.9, it/s 0.7
  6882. [0130 18:05:07 @multigpu.py:323] [p0574]  step: count(1463), step_time 1681.88, mean_step_time 1435.06, it/s 0.7
  6883. [0130 18:05:09 @multigpu.py:323] [p0574]  step: count(1464), step_time 1426.25, mean_step_time 1435.53, it/s 0.7
  6884. [0130 18:05:09 @multigpu.py:323] [p0115]  step: count(1449), step_time 1460.03, mean_step_time 1435.87, it/s 0.7
  6885. [0130 18:05:09 @multigpu.py:323] [p0576]  step: count(1473), step_time 1517.29, mean_step_time 1438.31, it/s 0.7
  6886. [0130 18:05:10 @multigpu.py:323] [p0574]  step: count(1465), step_time 1445.47, mean_step_time 1436.38, it/s 0.7
  6887. [0130 18:05:10 @multigpu.py:323] [p0115]  step: count(1450), step_time 1445.33, mean_step_time 1435.6, it/s 0.7
  6888. [0130 18:05:10 @multigpu.py:323] [p0576]  step: count(1474), step_time 1437.59, mean_step_time 1438.41, it/s 0.7
  6889. [0130 18:05:12 @multigpu.py:323] [p0574]  step: count(1466), step_time 1384.36, mean_step_time 1435.74, it/s 0.7
  6890. [0130 18:05:12 @multigpu.py:323] [p0115]  step: count(1451), step_time 1414.79, mean_step_time 1434.87, it/s 0.7
  6891. [0130 18:05:12 @multigpu.py:323] [p0576]  step: count(1475), step_time 1501.42, mean_step_time 1441.19, it/s 0.69
  6892. [0130 18:05:13 @multigpu.py:323] [p0574]  step: count(1467), step_time 1430.6, mean_step_time 1436.8, it/s 0.7
  6893. [0130 18:05:13 @multigpu.py:323] [p0115]  step: count(1452), step_time 1469.14, mean_step_time 1435.96, it/s 0.7
  6894. [0130 18:05:13 @multigpu.py:323] [p0576]  step: count(1476), step_time 1433.88, mean_step_time 1441.82, it/s 0.69
  6895. [0130 18:05:14 @multigpu.py:323] [p0574]  step: count(1468), step_time 1419.41, mean_step_time 1437.51, it/s 0.7
  6896. [0130 18:05:15 @multigpu.py:323] [p0115]  step: count(1453), step_time 1404.76, mean_step_time 1434.94, it/s 0.7
  6897. [0130 18:05:15 @multigpu.py:323] [p0576]  step: count(1477), step_time 1375.04, mean_step_time 1439.71, it/s 0.69
  6898. [0130 18:05:16 @multigpu.py:323] [p0574]  step: count(1469), step_time 1436.21, mean_step_time 1438.56, it/s 0.7
  6899. [0130 18:05:16 @multigpu.py:323] [p0115]  step: count(1454), step_time 1409.09, mean_step_time 1435.4, it/s 0.7
  6900. [0130 18:05:16 @multigpu.py:323] [p0576]  step: count(1478), step_time 1437.64, mean_step_time 1439.12, it/s 0.69
  6901. [0130 18:05:17 @multigpu.py:323] [p0574]  step: count(1470), step_time 1467.17, mean_step_time 1437.73, it/s 0.7
  6902. [0130 18:05:17 @multigpu.py:323] [p0115]  step: count(1455), step_time 1412.25, mean_step_time 1435.12, it/s 0.7
  6903. [0130 18:05:17 @multigpu.py:323] [p0576]  step: count(1479), step_time 1377.61, mean_step_time 1436.2, it/s 0.7
  6904. [0130 18:05:19 @multigpu.py:323] [p0574]  step: count(1471), step_time 1400.23, mean_step_time 1437.51, it/s 0.7
  6905. [0130 18:05:19 @multigpu.py:323] [p0115]  step: count(1456), step_time 1438.36, mean_step_time 1436.5, it/s 0.7
  6906. [0130 18:05:19 @multigpu.py:323] [p0576]  step: count(1480), step_time 1420.01, mean_step_time 1436.4, it/s 0.7
  6907. [0130 18:05:20 @multigpu.py:323] [p0115]  step: count(1457), step_time 1414.06, mean_step_time 1434.26, it/s 0.7
  6908. [0130 18:05:20 @multigpu.py:323] [p0574]  step: count(1472), step_time 1526.73, mean_step_time 1441.7, it/s 0.69
  6909. [0130 18:05:20 @multigpu.py:323] [p0576]  step: count(1481), step_time 1460.71, mean_step_time 1439.6, it/s 0.69
  6910. [0130 18:05:22 @multigpu.py:323] [p0115]  step: count(1458), step_time 1444.06, mean_step_time 1434.73, it/s 0.7
  6911. [0130 18:05:22 @multigpu.py:323] [p0576]  step: count(1482), step_time 1396.01, mean_step_time 1438.71, it/s 0.7
  6912. [0130 18:05:22 @multigpu.py:323] [p0574]  step: count(1473), step_time 1476.44, mean_step_time 1439.58, it/s 0.69
  6913. [0130 18:05:23 @multigpu.py:323] [p0115]  step: count(1459), step_time 1415.03, mean_step_time 1431.76, it/s 0.7
  6914. [0130 18:05:23 @multigpu.py:323] [p0576]  step: count(1483), step_time 1455.79, mean_step_time 1442.36, it/s 0.69
  6915. [0130 18:05:23 @multigpu.py:323] [p0574]  step: count(1474), step_time 1441.3, mean_step_time 1442.55, it/s 0.69
  6916. [0130 18:05:24 @multigpu.py:323] [p0115]  step: count(1460), step_time 1423.93, mean_step_time 1431.39, it/s 0.7
  6917. [0130 18:05:25 @multigpu.py:323] [p0576]  step: count(1484), step_time 1390.01, mean_step_time 1439.54, it/s 0.69
  6918. [0130 18:05:25 @multigpu.py:323] [p0574]  step: count(1475), step_time 1404.97, mean_step_time 1440.9, it/s 0.69
  6919. [0130 18:05:26 @multigpu.py:323] [p0115]  step: count(1461), step_time 1440.58, mean_step_time 1430.61, it/s 0.7
  6920. [0130 18:05:26 @multigpu.py:323] [p0576]  step: count(1485), step_time 1408.88, mean_step_time 1440.16, it/s 0.69
  6921. [0130 18:05:26 @multigpu.py:323] [p0574]  step: count(1476), step_time 1362.87, mean_step_time 1440.72, it/s 0.69
  6922. sending to address tcp://p0112:61216
  6923. ##### Sending to neptune:  online_score :  0.590532777773 , 1.6 #####
  6924. [u'online', 1.6]
  6925. receiving
  6926. [0130 18:05:27 @multigpu.py:323] [p0115]  step: count(1462), step_time 1486.69, mean_step_time 1432.32, it/s 0.7
  6927. [0130 18:05:27 @multigpu.py:323] [p0576]  step: count(1486), step_time 1470.78, mean_step_time 1443.38, it/s 0.69
  6928. [0130 18:05:27 @multigpu.py:323] [p0574]  step: count(1477), step_time 1481.01, mean_step_time 1443.34, it/s 0.69
  6929. sending to address tcp://p0112:61216
  6930. ##### Sending to neptune:  online_score :  0.59100319412 , 2.0 #####
  6931. [u'online', 2.0]
  6932. receiving
  6933. [0130 18:05:29 @multigpu.py:323] [p0576]  step: count(1487), step_time 1376.85, mean_step_time 1437.92, it/s 0.7
  6934. [0130 18:05:29 @multigpu.py:323] [p0115]  step: count(1463), step_time 1385.96, mean_step_time 1430.75, it/s 0.7
  6935. [0130 18:05:29 @multigpu.py:323] [p0574]  step: count(1478), step_time 1408.74, mean_step_time 1442.33, it/s 0.69
  6936. [0130 18:05:30 @multigpu.py:323] [p0576]  step: count(1488), step_time 1440.22, mean_step_time 1437.79, it/s 0.7
  6937. [0130 18:05:30 @multigpu.py:323] [p0574]  step: count(1479), step_time 1449.09, mean_step_time 1441.75, it/s 0.69
  6938. [0130 18:05:30 @multigpu.py:323] [p0115]  step: count(1464), step_time 1520.66, mean_step_time 1434.59, it/s 0.7
  6939. [0130 18:05:32 @multigpu.py:323] [p0574]  step: count(1480), step_time 1359.54, mean_step_time 1441.86, it/s 0.69
  6940. [0130 18:05:32 @multigpu.py:323] [p0576]  step: count(1489), step_time 1426.31, mean_step_time 1438.02, it/s 0.7
  6941. [0130 18:05:32 @multigpu.py:323] [p0115]  step: count(1465), step_time 1408.79, mean_step_time 1435.5, it/s 0.7
  6942. sending to address tcp://p0112:61216
  6943. ##### Sending to neptune:  online_score :  0.592168379691 , 1.3 #####
  6944. [u'online', 1.3]
  6945. ##### Sending to neptune:  active_workers :  0.592168413864 , 3 #####
  6946. receiving
  6947. [0130 18:05:33 @multigpu.py:323] [p0576]  step: count(1490), step_time 1440.04, mean_step_time 1440.13, it/s 0.69
  6948. [0130 18:05:33 @multigpu.py:323] [p0574]  step: count(1481), step_time 1445.72, mean_step_time 1442.19, it/s 0.69
  6949. [0130 18:05:33 @multigpu.py:323] [p0115]  step: count(1466), step_time 1403.25, mean_step_time 1432.71, it/s 0.7
  6950. [0130 18:05:35 @multigpu.py:323] [p0574]  step: count(1482), step_time 1407.35, mean_step_time 1442.77, it/s 0.69
  6951. [0130 18:05:35 @multigpu.py:323] [p0576]  step: count(1491), step_time 1422.11, mean_step_time 1436.68, it/s 0.7
  6952. [0130 18:05:35 @multigpu.py:323] [p0115]  step: count(1467), step_time 1445.25, mean_step_time 1432.91, it/s 0.7
  6953. [0130 18:05:36 @multigpu.py:323] [p0115]  step: count(1468), step_time 1456.31, mean_step_time 1434.92, it/s 0.7
  6954. [0130 18:05:36 @multigpu.py:323] [p0574]  step: count(1483), step_time 1523.84, mean_step_time 1434.87, it/s 0.7
  6955. [0130 18:05:36 @multigpu.py:323] [p0576]  step: count(1492), step_time 1508.82, mean_step_time 1434.85, it/s 0.7
  6956. [0130 18:05:37 @multigpu.py:323] [p0574]  step: count(1484), step_time 1402.04, mean_step_time 1433.66, it/s 0.7
  6957. [0130 18:05:37 @multigpu.py:323] [p0115]  step: count(1469), step_time 1415.48, mean_step_time 1432.69, it/s 0.7
  6958. [0130 18:05:37 @multigpu.py:323] [p0576]  step: count(1493), step_time 1444.5, mean_step_time 1431.21, it/s 0.7
  6959. [0130 18:05:39 @multigpu.py:323] [p0576]  step: count(1494), step_time 1410.45, mean_step_time 1429.85, it/s 0.7
  6960. [0130 18:05:39 @multigpu.py:323] [p0115]  step: count(1470), step_time 1446.5, mean_step_time 1432.75, it/s 0.7
  6961. [0130 18:05:39 @multigpu.py:323] [p0574]  step: count(1485), step_time 1516.5, mean_step_time 1437.21, it/s 0.7
  6962. [0130 18:05:40 @multigpu.py:323] [p0115]  step: count(1471), step_time 1382.08, mean_step_time 1431.11, it/s 0.7
  6963. [0130 18:05:40 @multigpu.py:323] [p0576]  step: count(1495), step_time 1417.69, mean_step_time 1425.67, it/s 0.7
  6964. [0130 18:05:40 @multigpu.py:323] [p0574]  step: count(1486), step_time 1456.4, mean_step_time 1440.81, it/s 0.69
  6965. [0130 18:05:42 @multigpu.py:323] [p0115]  step: count(1472), step_time 1440.42, mean_step_time 1429.68, it/s 0.7
  6966. [0130 18:05:42 @multigpu.py:323] [p0576]  step: count(1496), step_time 1468.62, mean_step_time 1427.4, it/s 0.7
  6967. [0130 18:05:42 @multigpu.py:323] [p0574]  step: count(1487), step_time 1463.15, mean_step_time 1442.44, it/s 0.69
  6968. [0130 18:05:43 @multigpu.py:323] [p0115]  step: count(1473), step_time 1427.59, mean_step_time 1430.82, it/s 0.7
  6969. [0130 18:05:43 @multigpu.py:323] [p0576]  step: count(1497), step_time 1416.67, mean_step_time 1429.49, it/s 0.7
  6970. [0130 18:05:43 @multigpu.py:323] [p0574]  step: count(1488), step_time 1377.04, mean_step_time 1440.32, it/s 0.69
  6971. [0130 18:05:45 @multigpu.py:323] [p0115]  step: count(1474), step_time 1389.29, mean_step_time 1429.83, it/s 0.7
  6972. [0130 18:05:45 @multigpu.py:323] [p0576]  step: count(1498), step_time 1364.5, mean_step_time 1425.83, it/s 0.7
  6973. [0130 18:05:45 @multigpu.py:323] [p0574]  step: count(1489), step_time 1432.0, mean_step_time 1440.11, it/s 0.69
  6974. [0130 18:05:46 @multigpu.py:323] [p0115]  step: count(1475), step_time 1449.54, mean_step_time 1431.69, it/s 0.7
  6975. [0130 18:05:46 @multigpu.py:323] [p0576]  step: count(1499), step_time 1459.3, mean_step_time 1429.91, it/s 0.7
  6976. [0130 18:05:46 @multigpu.py:323] [p0574]  step: count(1490), step_time 1426.07, mean_step_time 1438.05, it/s 0.7
  6977. [0130 18:05:47 @multigpu.py:323] [p0115]  step: count(1476), step_time 1472.42, mean_step_time 1433.4, it/s 0.7
  6978. [0130 18:05:47 @multigpu.py:323] [p0576]  step: count(1500), step_time 1443.76, mean_step_time 1431.1, it/s 0.7
  6979. sending debugging info...
  6980. sending to address tcp://p0112:61216
  6981. ##### Sending to neptune:  mean_delay :  0.596350740261 , 0.0 #####
  6982. sending to address tcp://p0112:61216
  6983. ##### Sending to neptune:  max_delay :  0.596350740261 , -0.0 #####
  6984. ##### Sending to neptune:  min_delay :  0.596350740261 , -0.0 #####
  6985. [u'delays', [0.0, -0.0, -0.0]]
  6986. receiving
  6987. ##### Sending to neptune:  cost :  0.596351259947 , -0.00888612959534 #####
  6988. sending to address tcp://p0112:61216
  6989. ##### Sending to neptune:  policy_loss :  0.596351259947 , -0.356634289026 #####
  6990. ##### Sending to neptune:  xentropy_loss :  0.596351259947 , -2.28363466263 #####
  6991. ##### Sending to neptune:  value_loss :  0.596351259947 , 1.50284433365 #####
  6992. ##### Sending to neptune:  advantage :  0.596351259947 , 0.00149029435124 #####
  6993. ##### Sending to neptune:  pred_reward :  0.596351259947 , 0.461806833744 #####
  6994. ##### Sending to neptune:  max_logit :  0.596351259947 , 0.228477463126 #####
  6995. [u'loss', -0.008886129595339298, -0.3566342890262604, -2.283634662628174, 1.5028443336486816, 0.0014902943512424827, 0.4618068337440491, 0.22847746312618256]
  6996. receiving
  6997. ##### Sending to neptune:  active_relus :  0.596351694663 , 9717197.02 #####
  6998. ##### Sending to neptune:  dp_per_s :  0.596351694663 , 89.2814797357 #####
  6999. [u'other', 9717197.02, 89.28147973567819]
  7000. receiving
  7001. [0130 18:05:48 @multigpu.py:323] [p0574]  step: count(1491), step_time 1440.54, mean_step_time 1440.07, it/s 0.69
  7002. [0130 18:05:49 @multigpu.py:323] [p0115]  step: count(1477), step_time 1435.04, mean_step_time 1434.44, it/s 0.7
  7003. [0130 18:05:49 @multigpu.py:323] [p0576]  step: count(1501), step_time 1432.08, mean_step_time 1429.67, it/s 0.7
  7004. [0130 18:05:49 @multigpu.py:323] [p0574]  step: count(1492), step_time 1429.17, mean_step_time 1435.19, it/s 0.7
  7005. [0130 18:05:50 @multigpu.py:323] [p0576]  step: count(1502), step_time 1436.44, mean_step_time 1431.69, it/s 0.7
  7006. [0130 18:05:50 @multigpu.py:323] [p0115]  step: count(1478), step_time 1457.75, mean_step_time 1435.13, it/s 0.7
  7007. [0130 18:05:50 @multigpu.py:323] [p0574]  step: count(1493), step_time 1425.53, mean_step_time 1432.64, it/s 0.7
  7008. [0130 18:05:52 @multigpu.py:323] [p0115]  step: count(1479), step_time 1425.16, mean_step_time 1435.64, it/s 0.7
  7009. [0130 18:05:52 @multigpu.py:323] [p0576]  step: count(1503), step_time 1450.49, mean_step_time 1431.43, it/s 0.7
  7010. [0130 18:05:52 @multigpu.py:323] [p0574]  step: count(1494), step_time 1463.16, mean_step_time 1433.74, it/s 0.7
  7011. [0130 18:05:53 @multigpu.py:323] [p0576]  step: count(1504), step_time 1405.55, mean_step_time 1432.2, it/s 0.7
  7012. [0130 18:05:53 @multigpu.py:323] [p0115]  step: count(1480), step_time 1427.44, mean_step_time 1435.81, it/s 0.7
  7013. [0130 18:05:53 @multigpu.py:323] [p0574]  step: count(1495), step_time 1400.87, mean_step_time 1433.53, it/s 0.7
  7014. [0130 18:05:55 @multigpu.py:323] [p0576]  step: count(1505), step_time 1428.03, mean_step_time 1433.16, it/s 0.7
  7015. [0130 18:05:55 @multigpu.py:323] [p0115]  step: count(1481), step_time 1469.23, mean_step_time 1437.24, it/s 0.7
  7016. [0130 18:05:55 @multigpu.py:323] [p0574]  step: count(1496), step_time 1409.19, mean_step_time 1435.85, it/s 0.7
  7017. [0130 18:05:56 @multigpu.py:323] [p0576]  step: count(1506), step_time 1411.85, mean_step_time 1430.21, it/s 0.7
  7018. [0130 18:05:56 @multigpu.py:323] [p0574]  step: count(1497), step_time 1441.86, mean_step_time 1433.89, it/s 0.7
  7019. [0130 18:05:56 @multigpu.py:323] [p0115]  step: count(1482), step_time 1455.18, mean_step_time 1435.67, it/s 0.7
  7020. [0130 18:05:57 @multigpu.py:323] [p0576]  step: count(1507), step_time 1409.49, mean_step_time 1431.85, it/s 0.7
  7021. [0130 18:05:58 @multigpu.py:323] [p0574]  step: count(1498), step_time 1387.26, mean_step_time 1432.82, it/s 0.7
  7022. [0130 18:05:58 @multigpu.py:323] [p0115]  step: count(1483), step_time 1448.08, mean_step_time 1438.77, it/s 0.7
  7023. sending to address tcp://p0112:61216
  7024. ##### Sending to neptune:  online_score :  0.599309429129 , 1.8 #####
  7025. [u'online', 1.8]
  7026. receiving
  7027. [0130 18:05:59 @multigpu.py:323] [p0576]  step: count(1508), step_time 1413.71, mean_step_time 1430.52, it/s 0.7
  7028. [0130 18:05:59 @multigpu.py:323] [p0574]  step: count(1499), step_time 1474.19, mean_step_time 1434.07, it/s 0.7
  7029. [0130 18:05:59 @multigpu.py:323] [p0115]  step: count(1484), step_time 1407.54, mean_step_time 1433.12, it/s 0.7
  7030. sending to address tcp://p0112:61216
  7031. ##### Sending to neptune:  online_score :  0.599584539665 , 1.7 #####
  7032. [u'online', 1.7]
  7033. receiving
  7034. sending to address tcp://p0112:61216
  7035. ##### Sending to neptune:  online_score :  0.599748105274 , 1.6 #####
  7036. [u'online', 1.6]
  7037. receiving
  7038. [0130 18:06:00 @multigpu.py:323] [p0576]  step: count(1509), step_time 1433.24, mean_step_time 1430.87, it/s 0.7
  7039. [0130 18:06:00 @multigpu.py:323] [p0574]  step: count(1500), step_time 1422.44, mean_step_time 1437.22, it/s 0.7
  7040. [0130 18:06:00 @multigpu.py:323] [p0115]  step: count(1485), step_time 1418.1, mean_step_time 1433.58, it/s 0.7
  7041. sending debugging info...
  7042. sending to address tcp://p0112:61216
  7043. ##### Sending to neptune:  mean_delay :  0.599946719408 , 0.0 #####
  7044. sending to address tcp://p0112:61216
  7045. ##### Sending to neptune:  max_delay :  0.599946719408 , -0.0 #####
  7046. ##### Sending to neptune:  min_delay :  0.599946719408 , -0.0 #####
  7047. [u'delays', [0.0, -0.0, -0.0]]
  7048. receiving
  7049. ##### Sending to neptune:  cost :  0.599947186046 , -0.00632883608341 #####
  7050. sending to address tcp://p0112:61216
  7051. ##### Sending to neptune:  policy_loss :  0.599947186046 , -0.0338481925428 #####
  7052. ##### Sending to neptune:  xentropy_loss :  0.599947186046 , -2.28367400169 #####
  7053. ##### Sending to neptune:  value_loss :  0.599947186046 , 1.50743150711 #####
  7054. ##### Sending to neptune:  advantage :  0.599947186046 , 5.52561978111e-05 #####
  7055. ##### Sending to neptune:  pred_reward :  0.599947186046 , 0.468498200178 #####
  7056. ##### Sending to neptune:  max_logit :  0.599947186046 , 0.22827039659 #####
  7057. [u'loss', -0.00632883608341217, -0.03384819254279137, -2.2836740016937256, 1.5074315071105957, 5.525619781110436e-05, 0.46849820017814636, 0.22827039659023285]
  7058. receiving
  7059. ##### Sending to neptune:  active_relus :  0.599947664407 , 9739097.15 #####
  7060. ##### Sending to neptune:  dp_per_s :  0.599947664407 , 89.2479513768 #####
  7061. [u'other', 9739097.15, 89.24795137675618]
  7062. receiving
  7063. [0130 18:06:02 @multigpu.py:323] [p0576]  step: count(1510), step_time 1447.74, mean_step_time 1431.25, it/s 0.7
  7064. [0130 18:06:02 @multigpu.py:323] [p0115]  step: count(1486), step_time 1410.68, mean_step_time 1433.95, it/s 0.7
  7065. [0130 18:06:02 @multigpu.py:323] [p0574]  step: count(1501), step_time 1438.43, mean_step_time 1436.85, it/s 0.7
  7066. [0130 18:06:03 @multigpu.py:323] [p0576]  step: count(1511), step_time 1375.44, mean_step_time 1428.92, it/s 0.7
  7067. [0130 18:06:03 @multigpu.py:323] [p0115]  step: count(1487), step_time 1439.88, mean_step_time 1433.69, it/s 0.7
  7068. [0130 18:06:03 @multigpu.py:323] [p0574]  step: count(1502), step_time 1416.34, mean_step_time 1437.3, it/s 0.7
  7069. [0130 18:06:05 @multigpu.py:323] [p0115]  step: count(1488), step_time 1430.65, mean_step_time 1432.4, it/s 0.7
  7070. [0130 18:06:05 @multigpu.py:323] [p0574]  step: count(1503), step_time 1426.86, mean_step_time 1432.45, it/s 0.7
  7071. [0130 18:06:05 @multigpu.py:323] [p0576]  step: count(1512), step_time 1578.32, mean_step_time 1432.39, it/s 0.7
  7072. [0130 18:06:06 @multigpu.py:323] [p0574]  step: count(1504), step_time 1401.97, mean_step_time 1432.45, it/s 0.7
  7073. [0130 18:06:06 @multigpu.py:323] [p0115]  step: count(1489), step_time 1416.3, mean_step_time 1432.44, it/s 0.7
  7074. [0130 18:06:06 @multigpu.py:323] [p0576]  step: count(1513), step_time 1443.97, mean_step_time 1432.37, it/s 0.7
  7075. [0130 18:06:08 @multigpu.py:323] [p0576]  step: count(1514), step_time 1383.8, mean_step_time 1431.03, it/s 0.7
  7076. [0130 18:06:08 @multigpu.py:323] [p0574]  step: count(1505), step_time 1493.87, mean_step_time 1431.32, it/s 0.7
  7077. [0130 18:06:08 @multigpu.py:323] [p0115]  step: count(1490), step_time 1512.18, mean_step_time 1435.73, it/s 0.7
  7078. [0130 18:06:09 @multigpu.py:323] [p0576]  step: count(1515), step_time 1453.69, mean_step_time 1432.83, it/s 0.7
  7079. [0130 18:06:09 @multigpu.py:323] [p0115]  step: count(1491), step_time 1375.9, mean_step_time 1435.42, it/s 0.7
  7080. [0130 18:06:09 @multigpu.py:323] [p0574]  step: count(1506), step_time 1456.25, mean_step_time 1431.31, it/s 0.7
  7081. [0130 18:06:10 @multigpu.py:323] [p0576]  step: count(1516), step_time 1394.5, mean_step_time 1429.13, it/s 0.7
  7082. [0130 18:06:10 @multigpu.py:323] [p0574]  step: count(1507), step_time 1372.31, mean_step_time 1426.77, it/s 0.7
  7083. [0130 18:06:10 @multigpu.py:323] [p0115]  step: count(1492), step_time 1453.23, mean_step_time 1436.06, it/s 0.7
  7084. [0130 18:06:12 @multigpu.py:323] [p0574]  step: count(1508), step_time 1397.08, mean_step_time 1427.77, it/s 0.7
  7085. [0130 18:06:12 @multigpu.py:323] [p0576]  step: count(1517), step_time 1456.13, mean_step_time 1431.1, it/s 0.7
  7086. [0130 18:06:12 @multigpu.py:323] [p0115]  step: count(1493), step_time 1446.72, mean_step_time 1437.02, it/s 0.7
  7087. [0130 18:06:13 @multigpu.py:323] [p0576]  step: count(1518), step_time 1364.42, mean_step_time 1431.1, it/s 0.7
  7088. [0130 18:06:13 @multigpu.py:323] [p0574]  step: count(1509), step_time 1491.4, mean_step_time 1430.74, it/s 0.7
  7089. [0130 18:06:13 @multigpu.py:323] [p0115]  step: count(1494), step_time 1462.95, mean_step_time 1440.7, it/s 0.69
  7090. [0130 18:06:15 @multigpu.py:323] [p0576]  step: count(1519), step_time 1408.41, mean_step_time 1428.55, it/s 0.7
  7091. [0130 18:06:15 @multigpu.py:323] [p0574]  step: count(1510), step_time 1446.65, mean_step_time 1431.77, it/s 0.7
  7092. [0130 18:06:15 @multigpu.py:323] [p0115]  step: count(1495), step_time 1475.53, mean_step_time 1442.0, it/s 0.69
  7093. [0130 18:06:16 @multigpu.py:323] [p0576]  step: count(1520), step_time 1400.12, mean_step_time 1426.37, it/s 0.7
  7094. [0130 18:06:16 @multigpu.py:323] [p0574]  step: count(1511), step_time 1419.34, mean_step_time 1430.71, it/s 0.7
  7095. [0130 18:06:16 @multigpu.py:323] [p0115]  step: count(1496), step_time 1459.53, mean_step_time 1441.35, it/s 0.69
  7096. [0130 18:06:17 @multigpu.py:323] [p0576]  step: count(1521), step_time 1461.32, mean_step_time 1427.83, it/s 0.7
  7097. [0130 18:06:18 @multigpu.py:323] [p0574]  step: count(1512), step_time 1409.34, mean_step_time 1429.72, it/s 0.7
  7098. [0130 18:06:18 @multigpu.py:323] [p0115]  step: count(1497), step_time 1461.02, mean_step_time 1442.65, it/s 0.69
  7099. [0130 18:06:19 @multigpu.py:323] [p0576]  step: count(1522), step_time 1395.16, mean_step_time 1425.77, it/s 0.7
  7100. [0130 18:06:19 @multigpu.py:323] [p0574]  step: count(1513), step_time 1454.11, mean_step_time 1431.14, it/s 0.7
  7101. [0130 18:06:19 @multigpu.py:323] [p0115]  step: count(1498), step_time 1485.19, mean_step_time 1444.02, it/s 0.69
  7102. [0130 18:06:20 @multigpu.py:323] [p0576]  step: count(1523), step_time 1449.23, mean_step_time 1425.71, it/s 0.7
  7103. [0130 18:06:20 @multigpu.py:323] [p0574]  step: count(1514), step_time 1400.58, mean_step_time 1428.02, it/s 0.7
  7104. [0130 18:06:21 @multigpu.py:323] [p0115]  step: count(1499), step_time 1429.95, mean_step_time 1444.26, it/s 0.69
  7105. [0130 18:06:22 @multigpu.py:323] [p0576]  step: count(1524), step_time 1418.53, mean_step_time 1426.35, it/s 0.7
  7106. [0130 18:06:22 @multigpu.py:323] [p0574]  step: count(1515), step_time 1416.47, mean_step_time 1428.8, it/s 0.7
  7107. [0130 18:06:22 @multigpu.py:323] [p0115]  step: count(1500), step_time 1381.96, mean_step_time 1441.99, it/s 0.69
  7108. sending debugging info...
  7109. sending to address tcp://p0112:61216
  7110. ##### Sending to neptune:  mean_delay :  0.605960784952 , 0.0 #####
  7111. sending to address tcp://p0112:61216
  7112. ##### Sending to neptune:  max_delay :  0.605960784952 , -0.0 #####
  7113. ##### Sending to neptune:  min_delay :  0.605960784952 , -0.0 #####
  7114. [u'delays', [0.0, -0.0, -0.0]]
  7115. receiving
  7116. ##### Sending to neptune:  cost :  0.605961283313 , -0.00497105671093 #####
  7117. sending to address tcp://p0112:61216
  7118. ##### Sending to neptune:  policy_loss :  0.605961283313 , 0.0196081455797 #####
  7119. ##### Sending to neptune:  xentropy_loss :  0.605961283313 , -2.28358197212 #####
  7120. ##### Sending to neptune:  value_loss :  0.605961283313 , 1.62767851353 #####
  7121. ##### Sending to neptune:  advantage :  0.605961283313 , 0.000106235434941 #####
  7122. ##### Sending to neptune:  pred_reward :  0.605961283313 , 0.468377023935 #####
  7123. ##### Sending to neptune:  max_logit :  0.605961283313 , 0.224669262767 #####
  7124. [u'loss', -0.0049710567109286785, 0.0196081455796957, -2.2835819721221924, 1.6276785135269165, 0.0001062354349414818, 0.468377023935318, 0.22466926276683807]
  7125. receiving
  7126. ##### Sending to neptune:  active_relus :  0.605961754984 , 9702806.99 #####
  7127. ##### Sending to neptune:  dp_per_s :  0.605961754984 , 89.1716589292 #####
  7128. [u'other', 9702806.99, 89.17165892922594]
  7129. receiving
  7130. [0130 18:06:23 @multigpu.py:323] [p0576]  step: count(1525), step_time 1436.64, mean_step_time 1426.78, it/s 0.7
  7131. [0130 18:06:23 @multigpu.py:323] [p0574]  step: count(1516), step_time 1442.65, mean_step_time 1430.47, it/s 0.7
  7132. [0130 18:06:23 @multigpu.py:323] [p0115]  step: count(1501), step_time 1436.4, mean_step_time 1440.35, it/s 0.69
  7133. sending to address tcp://p0112:61216
  7134. ##### Sending to neptune:  online_score :  0.606569805278 , 2.3 #####
  7135. [u'online', 2.3]
  7136. receiving
  7137. [0130 18:06:25 @multigpu.py:323] [p0576]  step: count(1526), step_time 1410.51, mean_step_time 1426.72, it/s 0.7
  7138. [0130 18:06:25 @multigpu.py:323] [p0574]  step: count(1517), step_time 1453.84, mean_step_time 1431.07, it/s 0.7
  7139. [0130 18:06:25 @multigpu.py:323] [p0115]  step: count(1502), step_time 1457.57, mean_step_time 1440.47, it/s 0.69
  7140. sending to address tcp://p0112:61216
  7141. ##### Sending to neptune:  online_score :  0.606775739657 , 1.7 #####
  7142. [u'online', 1.7]
  7143. receiving
  7144. sending to address tcp://p0112:61216
  7145. ##### Sending to neptune:  online_score :  0.606916066607 , 2.2 #####
  7146. [u'online', 2.2]
  7147. receiving
  7148. [0130 18:06:26 @multigpu.py:323] [p0576]  step: count(1527), step_time 1413.27, mean_step_time 1426.91, it/s 0.7
  7149. [0130 18:06:26 @multigpu.py:323] [p0574]  step: count(1518), step_time 1464.12, mean_step_time 1434.91, it/s 0.7
  7150. [0130 18:06:26 @multigpu.py:323] [p0115]  step: count(1503), step_time 1385.83, mean_step_time 1437.36, it/s 0.7
  7151. [0130 18:06:27 @multigpu.py:323] [p0576]  step: count(1528), step_time 1397.73, mean_step_time 1426.11, it/s 0.7
  7152. [0130 18:06:28 @multigpu.py:323] [p0574]  step: count(1519), step_time 1409.67, mean_step_time 1431.69, it/s 0.7
  7153. [0130 18:06:28 @multigpu.py:323] [p0115]  step: count(1504), step_time 1424.17, mean_step_time 1438.19, it/s 0.7
  7154. [0130 18:06:29 @multigpu.py:323] [p0576]  step: count(1529), step_time 1454.14, mean_step_time 1427.15, it/s 0.7
  7155. [0130 18:06:29 @multigpu.py:323] [p0574]  step: count(1520), step_time 1500.47, mean_step_time 1435.59, it/s 0.7
  7156. [0130 18:06:29 @multigpu.py:323] [p0115]  step: count(1505), step_time 1486.96, mean_step_time 1441.63, it/s 0.69
  7157. [0130 18:06:30 @multigpu.py:323] [p0576]  step: count(1530), step_time 1396.72, mean_step_time 1424.6, it/s 0.7
  7158. [0130 18:06:31 @multigpu.py:323] [p0574]  step: count(1521), step_time 1421.62, mean_step_time 1434.75, it/s 0.7
  7159. [0130 18:06:31 @multigpu.py:323] [p0115]  step: count(1506), step_time 1443.23, mean_step_time 1443.26, it/s 0.69
  7160. [0130 18:06:32 @multigpu.py:323] [p0576]  step: count(1531), step_time 1410.17, mean_step_time 1426.34, it/s 0.7
  7161. [0130 18:06:32 @multigpu.py:323] [p0574]  step: count(1522), step_time 1412.52, mean_step_time 1434.56, it/s 0.7
  7162. [0130 18:06:32 @multigpu.py:323] [p0115]  step: count(1507), step_time 1414.24, mean_step_time 1441.98, it/s 0.69
  7163. [0130 18:06:34 @multigpu.py:323] [p0115]  step: count(1508), step_time 1425.03, mean_step_time 1441.69, it/s 0.69
  7164. [0130 18:06:34 @multigpu.py:323] [p0574]  step: count(1523), step_time 1528.19, mean_step_time 1439.62, it/s 0.69
  7165. [0130 18:06:34 @multigpu.py:323] [p0576]  step: count(1532), step_time 1884.68, mean_step_time 1441.66, it/s 0.69
  7166. [0130 18:06:35 @multigpu.py:323] [p0576]  step: count(1533), step_time 1429.69, mean_step_time 1440.94, it/s 0.69
  7167. [0130 18:06:35 @multigpu.py:323] [p0115]  step: count(1509), step_time 1459.08, mean_step_time 1443.83, it/s 0.69
  7168. [0130 18:06:35 @multigpu.py:323] [p0574]  step: count(1524), step_time 1493.5, mean_step_time 1444.2, it/s 0.69
  7169. [0130 18:06:36 @multigpu.py:323] [p0576]  step: count(1534), step_time 1403.82, mean_step_time 1441.94, it/s 0.69
  7170. [0130 18:06:36 @multigpu.py:323] [p0115]  step: count(1510), step_time 1406.96, mean_step_time 1438.57, it/s 0.7
  7171. [0130 18:06:36 @multigpu.py:323] [p0574]  step: count(1525), step_time 1449.17, mean_step_time 1441.96, it/s 0.69
  7172. [0130 18:06:38 @multigpu.py:323] [p0576]  step: count(1535), step_time 1401.52, mean_step_time 1439.34, it/s 0.69
  7173. [0130 18:06:38 @multigpu.py:323] [p0115]  step: count(1511), step_time 1409.91, mean_step_time 1440.27, it/s 0.69
  7174. [0130 18:06:38 @multigpu.py:323] [p0574]  step: count(1526), step_time 1434.71, mean_step_time 1440.89, it/s 0.69
  7175. [0130 18:06:39 @multigpu.py:323] [p0115]  step: count(1512), step_time 1424.66, mean_step_time 1438.84, it/s 0.7
  7176. [0130 18:06:39 @multigpu.py:323] [p0576]  step: count(1536), step_time 1467.83, mean_step_time 1443.0, it/s 0.69
  7177. [0130 18:06:39 @multigpu.py:323] [p0574]  step: count(1527), step_time 1399.48, mean_step_time 1442.25, it/s 0.69
  7178. [0130 18:06:41 @multigpu.py:323] [p0115]  step: count(1513), step_time 1423.8, mean_step_time 1437.7, it/s 0.7
  7179. [0130 18:06:41 @multigpu.py:323] [p0576]  step: count(1537), step_time 1449.03, mean_step_time 1442.65, it/s 0.69
  7180. [0130 18:06:41 @multigpu.py:323] [p0574]  step: count(1528), step_time 1401.56, mean_step_time 1442.47, it/s 0.69
  7181. [0130 18:06:42 @multigpu.py:323] [p0115]  step: count(1514), step_time 1410.1, mean_step_time 1435.06, it/s 0.7
  7182. [0130 18:06:42 @multigpu.py:323] [p0576]  step: count(1538), step_time 1480.84, mean_step_time 1448.47, it/s 0.69
  7183. [0130 18:06:42 @multigpu.py:323] [p0574]  step: count(1529), step_time 1472.88, mean_step_time 1441.54, it/s 0.69
  7184. [0130 18:06:43 @multigpu.py:323] [p0115]  step: count(1515), step_time 1391.68, mean_step_time 1430.86, it/s 0.7
  7185. [0130 18:06:44 @multigpu.py:323] [p0574]  step: count(1530), step_time 1419.66, mean_step_time 1440.19, it/s 0.69
  7186. [0130 18:06:44 @multigpu.py:323] [p0576]  step: count(1539), step_time 1470.23, mean_step_time 1451.56, it/s 0.69
  7187. [0130 18:06:45 @multigpu.py:323] [p0115]  step: count(1516), step_time 1472.46, mean_step_time 1431.51, it/s 0.7
  7188. [0130 18:06:45 @multigpu.py:323] [p0576]  step: count(1540), step_time 1379.21, mean_step_time 1450.51, it/s 0.69
  7189. [0130 18:06:45 @multigpu.py:323] [p0574]  step: count(1531), step_time 1432.77, mean_step_time 1440.87, it/s 0.69
  7190. [0130 18:06:46 @multigpu.py:323] [p0115]  step: count(1517), step_time 1476.83, mean_step_time 1432.3, it/s 0.7
  7191. [0130 18:06:46 @multigpu.py:323] [p0574]  step: count(1532), step_time 1408.65, mean_step_time 1440.83, it/s 0.69
  7192. [0130 18:06:46 @multigpu.py:323] [p0576]  step: count(1541), step_time 1439.1, mean_step_time 1449.4, it/s 0.69
  7193. [0130 18:06:48 @multigpu.py:323] [p0115]  step: count(1518), step_time 1387.91, mean_step_time 1427.44, it/s 0.7
  7194. [0130 18:06:48 @multigpu.py:323] [p0576]  step: count(1542), step_time 1395.19, mean_step_time 1449.4, it/s 0.69
  7195. [0130 18:06:48 @multigpu.py:323] [p0574]  step: count(1533), step_time 1452.84, mean_step_time 1440.77, it/s 0.69
  7196. [0130 18:06:49 @multigpu.py:323] [p0576]  step: count(1543), step_time 1407.96, mean_step_time 1447.34, it/s 0.69
  7197. [0130 18:06:49 @multigpu.py:323] [p0115]  step: count(1519), step_time 1479.98, mean_step_time 1429.94, it/s 0.7
  7198. [0130 18:06:49 @multigpu.py:323] [p0574]  step: count(1534), step_time 1448.62, mean_step_time 1443.17, it/s 0.69
  7199. [0130 18:06:51 @multigpu.py:323] [p0576]  step: count(1544), step_time 1377.59, mean_step_time 1445.29, it/s 0.69
  7200. [0130 18:06:51 @multigpu.py:323] [p0115]  step: count(1520), step_time 1407.82, mean_step_time 1431.23, it/s 0.7
  7201. [0130 18:06:51 @multigpu.py:323] [p0574]  step: count(1535), step_time 1427.36, mean_step_time 1443.72, it/s 0.69
  7202. [0130 18:06:52 @multigpu.py:323] [p0576]  step: count(1545), step_time 1436.21, mean_step_time 1445.27, it/s 0.69
  7203. [0130 18:06:52 @multigpu.py:323] [p0115]  step: count(1521), step_time 1458.2, mean_step_time 1432.32, it/s 0.7
  7204. [0130 18:06:52 @multigpu.py:323] [p0574]  step: count(1536), step_time 1439.56, mean_step_time 1443.56, it/s 0.69
  7205. sending to address tcp://p0112:61216
  7206. ##### Sending to neptune:  online_score :  0.614415989121 , 0.8 #####
  7207. [u'online', 0.8]
  7208. ##### Sending to neptune:  active_workers :  0.614416100515 , 3 #####
  7209. receiving
  7210. sending to address tcp://p0112:61216
  7211. ##### Sending to neptune:  online_score :  0.614469838606 , 1.7 #####
  7212. [u'online', 1.7]
  7213. receiving
  7214. [0130 18:06:53 @multigpu.py:323] [p0576]  step: count(1546), step_time 1365.79, mean_step_time 1443.04, it/s 0.69
  7215. [0130 18:06:54 @multigpu.py:323] [p0115]  step: count(1522), step_time 1442.97, mean_step_time 1431.59, it/s 0.7
  7216. [0130 18:06:54 @multigpu.py:323] [p0574]  step: count(1537), step_time 1399.53, mean_step_time 1440.85, it/s 0.69
  7217. [0130 18:06:55 @multigpu.py:323] [p0576]  step: count(1547), step_time 1463.65, mean_step_time 1445.56, it/s 0.69
  7218. [0130 18:06:55 @multigpu.py:323] [p0115]  step: count(1523), step_time 1383.89, mean_step_time 1431.5, it/s 0.7
  7219. [0130 18:06:55 @multigpu.py:323] [p0574]  step: count(1538), step_time 1405.73, mean_step_time 1437.93, it/s 0.7
  7220. [0130 18:06:56 @multigpu.py:323] [p0576]  step: count(1548), step_time 1425.38, mean_step_time 1446.94, it/s 0.69
  7221. [0130 18:06:56 @multigpu.py:323] [p0115]  step: count(1524), step_time 1458.14, mean_step_time 1433.19, it/s 0.7
  7222. [0130 18:06:57 @multigpu.py:323] [p0574]  step: count(1539), step_time 1469.78, mean_step_time 1440.93, it/s 0.69
  7223. sending to address tcp://p0112:61216
  7224. ##### Sending to neptune:  online_score :  0.615593735774 , 0.8 #####
  7225. [u'online', 0.8]
  7226. receiving
  7227. [0130 18:06:58 @multigpu.py:323] [p0576]  step: count(1549), step_time 1398.88, mean_step_time 1444.17, it/s 0.69
  7228. [0130 18:06:58 @multigpu.py:323] [p0115]  step: count(1525), step_time 1391.89, mean_step_time 1428.44, it/s 0.7
  7229. [0130 18:06:58 @multigpu.py:323] [p0574]  step: count(1540), step_time 1427.82, mean_step_time 1437.3, it/s 0.7
  7230. [0130 18:06:59 @multigpu.py:323] [p0576]  step: count(1550), step_time 1448.98, mean_step_time 1446.79, it/s 0.69
  7231. [0130 18:06:59 @multigpu.py:323] [p0115]  step: count(1526), step_time 1389.32, mean_step_time 1425.74, it/s 0.7
  7232. [0130 18:06:59 @multigpu.py:323] [p0574]  step: count(1541), step_time 1421.4, mean_step_time 1437.29, it/s 0.7
  7233. [0130 18:07:01 @multigpu.py:323] [p0115]  step: count(1527), step_time 1414.52, mean_step_time 1425.76, it/s 0.7
  7234. [0130 18:07:01 @multigpu.py:323] [p0576]  step: count(1551), step_time 1488.68, mean_step_time 1450.71, it/s 0.69
  7235. [0130 18:07:01 @multigpu.py:323] [p0574]  step: count(1542), step_time 1428.02, mean_step_time 1438.06, it/s 0.7
  7236. [0130 18:07:02 @multigpu.py:323] [p0574]  step: count(1543), step_time 1433.98, mean_step_time 1433.35, it/s 0.7
  7237. [0130 18:07:02 @multigpu.py:323] [p0115]  step: count(1528), step_time 1574.99, mean_step_time 1433.26, it/s 0.7
  7238. [0130 18:07:02 @multigpu.py:323] [p0576]  step: count(1552), step_time 1535.72, mean_step_time 1433.27, it/s 0.7
  7239. [0130 18:07:04 @multigpu.py:323] [p0115]  step: count(1529), step_time 1408.31, mean_step_time 1430.72, it/s 0.7
  7240. [0130 18:07:04 @multigpu.py:323] [p0576]  step: count(1553), step_time 1426.85, mean_step_time 1433.12, it/s 0.7
  7241. [0130 18:07:04 @multigpu.py:323] [p0574]  step: count(1544), step_time 1444.66, mean_step_time 1430.91, it/s 0.7
  7242. [0130 18:07:05 @multigpu.py:323] [p0576]  step: count(1554), step_time 1411.79, mean_step_time 1433.52, it/s 0.7
  7243. [0130 18:07:05 @multigpu.py:323] [p0115]  step: count(1530), step_time 1433.44, mean_step_time 1432.04, it/s 0.7
  7244. [0130 18:07:05 @multigpu.py:323] [p0574]  step: count(1545), step_time 1416.43, mean_step_time 1429.27, it/s 0.7
  7245. [0130 18:07:06 @multigpu.py:323] [p0576]  step: count(1555), step_time 1407.51, mean_step_time 1433.82, it/s 0.7
  7246. [0130 18:07:06 @multigpu.py:323] [p0115]  step: count(1531), step_time 1427.83, mean_step_time 1432.94, it/s 0.7
  7247. [0130 18:07:07 @multigpu.py:323] [p0574]  step: count(1546), step_time 1441.64, mean_step_time 1429.62, it/s 0.7
  7248. [0130 18:07:08 @multigpu.py:323] [p0576]  step: count(1556), step_time 1427.75, mean_step_time 1431.82, it/s 0.7
  7249. [0130 18:07:08 @multigpu.py:323] [p0115]  step: count(1532), step_time 1468.38, mean_step_time 1435.12, it/s 0.7
  7250. [0130 18:07:08 @multigpu.py:323] [p0574]  step: count(1547), step_time 1443.35, mean_step_time 1431.81, it/s 0.7
  7251. [0130 18:07:09 @multigpu.py:323] [p0576]  step: count(1557), step_time 1440.46, mean_step_time 1431.39, it/s 0.7
  7252. [0130 18:07:09 @multigpu.py:323] [p0115]  step: count(1533), step_time 1405.02, mean_step_time 1434.19, it/s 0.7
  7253. [0130 18:07:09 @multigpu.py:323] [p0574]  step: count(1548), step_time 1479.32, mean_step_time 1435.7, it/s 0.7
  7254. [0130 18:07:11 @multigpu.py:323] [p0576]  step: count(1558), step_time 1434.45, mean_step_time 1429.07, it/s 0.7
  7255. [0130 18:07:11 @multigpu.py:323] [p0115]  step: count(1534), step_time 1436.15, mean_step_time 1435.49, it/s 0.7
  7256. [0130 18:07:11 @multigpu.py:323] [p0574]  step: count(1549), step_time 1454.5, mean_step_time 1434.78, it/s 0.7
  7257. [0130 18:07:12 @multigpu.py:323] [p0576]  step: count(1559), step_time 1427.64, mean_step_time 1426.94, it/s 0.7
  7258. [0130 18:07:12 @multigpu.py:323] [p0115]  step: count(1535), step_time 1461.25, mean_step_time 1438.97, it/s 0.69
  7259. [0130 18:07:12 @multigpu.py:323] [p0574]  step: count(1550), step_time 1408.36, mean_step_time 1434.22, it/s 0.7
  7260. [0130 18:07:14 @multigpu.py:323] [p0576]  step: count(1560), step_time 1428.16, mean_step_time 1429.39, it/s 0.7
  7261. [0130 18:07:14 @multigpu.py:323] [p0115]  step: count(1536), step_time 1409.0, mean_step_time 1435.79, it/s 0.7
  7262. [0130 18:07:14 @multigpu.py:323] [p0574]  step: count(1551), step_time 1414.37, mean_step_time 1433.3, it/s 0.7
  7263. [0130 18:07:15 @multigpu.py:323] [p0576]  step: count(1561), step_time 1409.87, mean_step_time 1427.93, it/s 0.7
  7264. [0130 18:07:15 @multigpu.py:323] [p0115]  step: count(1537), step_time 1457.82, mean_step_time 1434.84, it/s 0.7
  7265. [0130 18:07:15 @multigpu.py:323] [p0574]  step: count(1552), step_time 1425.49, mean_step_time 1434.14, it/s 0.7
  7266. [0130 18:07:16 @multigpu.py:323] [p0576]  step: count(1562), step_time 1441.37, mean_step_time 1430.23, it/s 0.7
  7267. [0130 18:07:17 @multigpu.py:323] [p0115]  step: count(1538), step_time 1399.1, mean_step_time 1435.4, it/s 0.7
  7268. [0130 18:07:17 @multigpu.py:323] [p0574]  step: count(1553), step_time 1392.19, mean_step_time 1431.11, it/s 0.7
  7269. [0130 18:07:18 @multigpu.py:323] [p0576]  step: count(1563), step_time 1375.9, mean_step_time 1428.63, it/s 0.7
  7270. [0130 18:07:18 @multigpu.py:323] [p0115]  step: count(1539), step_time 1396.81, mean_step_time 1431.24, it/s 0.7
  7271. [0130 18:07:18 @multigpu.py:323] [p0574]  step: count(1554), step_time 1443.34, mean_step_time 1430.84, it/s 0.7
  7272. [0130 18:07:19 @multigpu.py:323] [p0576]  step: count(1564), step_time 1417.54, mean_step_time 1430.63, it/s 0.7
  7273. [0130 18:07:19 @multigpu.py:323] [p0115]  step: count(1540), step_time 1436.86, mean_step_time 1432.7, it/s 0.7
  7274. [0130 18:07:19 @multigpu.py:323] [p0574]  step: count(1555), step_time 1447.6, mean_step_time 1431.85, it/s 0.7
  7275. sending to address tcp://p0112:61216
  7276. ##### Sending to neptune:  online_score :  0.622098054952 , 1.3 #####
  7277. [u'online', 1.3]
  7278. receiving
  7279. [0130 18:07:21 @multigpu.py:323] [p0576]  step: count(1565), step_time 1399.06, mean_step_time 1428.77, it/s 0.7
  7280. [0130 18:07:21 @multigpu.py:323] [p0115]  step: count(1541), step_time 1431.0, mean_step_time 1431.34, it/s 0.7
  7281. [0130 18:07:21 @multigpu.py:323] [p0574]  step: count(1556), step_time 1404.22, mean_step_time 1430.09, it/s 0.7
  7282. [0130 18:07:22 @multigpu.py:323] [p0576]  step: count(1566), step_time 1383.49, mean_step_time 1429.66, it/s 0.7
  7283. [0130 18:07:22 @multigpu.py:323] [p0574]  step: count(1557), step_time 1425.37, mean_step_time 1431.38, it/s 0.7
  7284. [0130 18:07:22 @multigpu.py:323] [p0115]  step: count(1542), step_time 1472.93, mean_step_time 1432.83, it/s 0.7
  7285. [0130 18:07:24 @multigpu.py:323] [p0576]  step: count(1567), step_time 1476.75, mean_step_time 1430.31, it/s 0.7
  7286. [0130 18:07:24 @multigpu.py:323] [p0115]  step: count(1543), step_time 1435.89, mean_step_time 1435.43, it/s 0.7
  7287. [0130 18:07:24 @multigpu.py:323] [p0574]  step: count(1558), step_time 1442.92, mean_step_time 1433.24, it/s 0.7
  7288. [0130 18:07:25 @multigpu.py:323] [p0576]  step: count(1568), step_time 1391.28, mean_step_time 1428.61, it/s 0.7
  7289. [0130 18:07:25 @multigpu.py:323] [p0574]  step: count(1559), step_time 1423.02, mean_step_time 1430.9, it/s 0.7
  7290. [0130 18:07:25 @multigpu.py:323] [p0115]  step: count(1544), step_time 1434.17, mean_step_time 1434.23, it/s 0.7
  7291. sending to address tcp://p0112:61216
  7292. ##### Sending to neptune:  online_score :  0.623693410556 , 1.3 #####
  7293. [u'online', 1.3]
  7294. receiving
  7295. [0130 18:07:26 @multigpu.py:323] [p0576]  step: count(1569), step_time 1386.62, mean_step_time 1427.99, it/s 0.7
  7296. [0130 18:07:27 @multigpu.py:323] [p0574]  step: count(1560), step_time 1398.7, mean_step_time 1429.44, it/s 0.7
  7297. [0130 18:07:27 @multigpu.py:323] [p0115]  step: count(1545), step_time 1414.45, mean_step_time 1435.36, it/s 0.7
  7298. sending to address tcp://p0112:61216
  7299. ##### Sending to neptune:  online_score :  0.623916657435 , 1.0 #####
  7300. [u'online', 1.0]
  7301. receiving
  7302. [0130 18:07:28 @multigpu.py:323] [p0576]  step: count(1570), step_time 1444.13, mean_step_time 1427.75, it/s 0.7
  7303. [0130 18:07:28 @multigpu.py:323] [p0574]  step: count(1561), step_time 1390.77, mean_step_time 1427.91, it/s 0.7
  7304. [0130 18:07:28 @multigpu.py:323] [p0115]  step: count(1546), step_time 1464.95, mean_step_time 1439.14, it/s 0.69
  7305. [0130 18:07:29 @multigpu.py:323] [p0576]  step: count(1571), step_time 1393.22, mean_step_time 1422.98, it/s 0.7
  7306. [0130 18:07:29 @multigpu.py:323] [p0574]  step: count(1562), step_time 1461.98, mean_step_time 1429.61, it/s 0.7
  7307. [0130 18:07:29 @multigpu.py:323] [p0115]  step: count(1547), step_time 1375.92, mean_step_time 1437.21, it/s 0.7
  7308. [0130 18:07:31 @multigpu.py:323] [p0574]  step: count(1563), step_time 1461.96, mean_step_time 1431.01, it/s 0.7
  7309. [0130 18:07:31 @multigpu.py:323] [p0115]  step: count(1548), step_time 1447.15, mean_step_time 1430.82, it/s 0.7
  7310. [0130 18:07:31 @multigpu.py:323] [p0576]  step: count(1572), step_time 1695.82, mean_step_time 1430.98, it/s 0.7
  7311. [0130 18:07:32 @multigpu.py:323] [p0574]  step: count(1564), step_time 1392.05, mean_step_time 1428.38, it/s 0.7
  7312. [0130 18:07:32 @multigpu.py:323] [p0115]  step: count(1549), step_time 1419.71, mean_step_time 1431.39, it/s 0.7
  7313. [0130 18:07:32 @multigpu.py:323] [p0576]  step: count(1573), step_time 1422.03, mean_step_time 1430.74, it/s 0.7
  7314. [0130 18:07:34 @multigpu.py:323] [p0576]  step: count(1574), step_time 1374.04, mean_step_time 1428.85, it/s 0.7
  7315. [0130 18:07:34 @multigpu.py:323] [p0574]  step: count(1565), step_time 1451.87, mean_step_time 1430.15, it/s 0.7
  7316. [0130 18:07:34 @multigpu.py:323] [p0115]  step: count(1550), step_time 1432.63, mean_step_time 1431.35, it/s 0.7
  7317. [0130 18:07:35 @multigpu.py:323] [p0576]  step: count(1575), step_time 1461.69, mean_step_time 1431.56, it/s 0.7
  7318. [0130 18:07:35 @multigpu.py:323] [p0574]  step: count(1566), step_time 1442.13, mean_step_time 1430.18, it/s 0.7
  7319. [0130 18:07:35 @multigpu.py:323] [p0115]  step: count(1551), step_time 1473.36, mean_step_time 1433.63, it/s 0.7
  7320. [0130 18:07:37 @multigpu.py:323] [p0576]  step: count(1576), step_time 1428.61, mean_step_time 1431.61, it/s 0.7
  7321. [0130 18:07:37 @multigpu.py:323] [p0574]  step: count(1567), step_time 1442.95, mean_step_time 1430.16, it/s 0.7
  7322. [0130 18:07:37 @multigpu.py:323] [p0115]  step: count(1552), step_time 1424.53, mean_step_time 1431.44, it/s 0.7
  7323. [0130 18:07:38 @multigpu.py:323] [p0576]  step: count(1577), step_time 1451.4, mean_step_time 1432.15, it/s 0.7
  7324. [0130 18:07:38 @multigpu.py:323] [p0115]  step: count(1553), step_time 1432.63, mean_step_time 1432.82, it/s 0.7
  7325. [0130 18:07:38 @multigpu.py:323] [p0574]  step: count(1568), step_time 1483.12, mean_step_time 1430.35, it/s 0.7
  7326. [0130 18:07:39 @multigpu.py:323] [p0576]  step: count(1578), step_time 1415.32, mean_step_time 1431.2, it/s 0.7
  7327. [0130 18:07:39 @multigpu.py:323] [p0115]  step: count(1554), step_time 1393.06, mean_step_time 1430.66, it/s 0.7
  7328. [0130 18:07:40 @multigpu.py:323] [p0574]  step: count(1569), step_time 1439.62, mean_step_time 1429.6, it/s 0.7
  7329. [0130 18:07:41 @multigpu.py:323] [p0115]  step: count(1555), step_time 1408.59, mean_step_time 1428.03, it/s 0.7
  7330. [0130 18:07:41 @multigpu.py:323] [p0576]  step: count(1579), step_time 1419.95, mean_step_time 1430.81, it/s 0.7
  7331. [0130 18:07:41 @multigpu.py:323] [p0574]  step: count(1570), step_time 1429.09, mean_step_time 1430.64, it/s 0.7
  7332. [0130 18:07:42 @multigpu.py:323] [p0576]  step: count(1580), step_time 1417.08, mean_step_time 1430.26, it/s 0.7
  7333. [0130 18:07:42 @multigpu.py:323] [p0115]  step: count(1556), step_time 1468.9, mean_step_time 1431.02, it/s 0.7
  7334. [0130 18:07:42 @multigpu.py:323] [p0574]  step: count(1571), step_time 1420.37, mean_step_time 1430.94, it/s 0.7
  7335. [0130 18:07:44 @multigpu.py:323] [p0576]  step: count(1581), step_time 1426.62, mean_step_time 1431.1, it/s 0.7
  7336. [0130 18:07:44 @multigpu.py:323] [p0115]  step: count(1557), step_time 1384.78, mean_step_time 1427.37, it/s 0.7
  7337. [0130 18:07:44 @multigpu.py:323] [p0574]  step: count(1572), step_time 1442.39, mean_step_time 1431.78, it/s 0.7
  7338. [0130 18:07:45 @multigpu.py:323] [p0115]  step: count(1558), step_time 1406.15, mean_step_time 1427.72, it/s 0.7
  7339. [0130 18:07:45 @multigpu.py:323] [p0576]  step: count(1582), step_time 1482.99, mean_step_time 1433.18, it/s 0.7
  7340. [0130 18:07:45 @multigpu.py:323] [p0574]  step: count(1573), step_time 1434.9, mean_step_time 1433.92, it/s 0.7
  7341. sending to address tcp://p0112:61216
  7342. ##### Sending to neptune:  online_score :  0.629087399708 , 2.1 #####
  7343. [u'online', 2.1]
  7344. receiving
  7345. [0130 18:07:47 @multigpu.py:323] [p0115]  step: count(1559), step_time 1429.28, mean_step_time 1429.35, it/s 0.7
  7346. [0130 18:07:47 @multigpu.py:323] [p0576]  step: count(1583), step_time 1467.46, mean_step_time 1437.75, it/s 0.7
  7347. [0130 18:07:47 @multigpu.py:323] [p0574]  step: count(1574), step_time 1448.98, mean_step_time 1434.2, it/s 0.7
  7348. [0130 18:07:48 @multigpu.py:323] [p0115]  step: count(1560), step_time 1374.07, mean_step_time 1426.21, it/s 0.7
  7349. [0130 18:07:48 @multigpu.py:323] [p0574]  step: count(1575), step_time 1368.49, mean_step_time 1430.24, it/s 0.7
  7350. [0130 18:07:48 @multigpu.py:323] [p0576]  step: count(1584), step_time 1464.55, mean_step_time 1440.1, it/s 0.69
  7351. sending to address tcp://p0112:61216
  7352. ##### Sending to neptune:  online_score :  0.630022522211 , 1.6 #####
  7353. [u'online', 1.6]
  7354. receiving
  7355. [0130 18:07:49 @multigpu.py:323] [p0115]  step: count(1561), step_time 1453.9, mean_step_time 1427.35, it/s 0.7
  7356. [0130 18:07:50 @multigpu.py:323] [p0574]  step: count(1576), step_time 1467.22, mean_step_time 1433.39, it/s 0.7
  7357. [0130 18:07:50 @multigpu.py:323] [p0576]  step: count(1585), step_time 1512.55, mean_step_time 1445.78, it/s 0.69
  7358. [0130 18:07:51 @multigpu.py:323] [p0115]  step: count(1562), step_time 1488.79, mean_step_time 1428.15, it/s 0.7
  7359. [0130 18:07:51 @multigpu.py:323] [p0574]  step: count(1577), step_time 1436.96, mean_step_time 1433.97, it/s 0.7
  7360. [0130 18:07:51 @multigpu.py:323] [p0576]  step: count(1586), step_time 1385.99, mean_step_time 1445.9, it/s 0.69
  7361. [0130 18:07:52 @multigpu.py:323] [p0115]  step: count(1563), step_time 1464.72, mean_step_time 1429.59, it/s 0.7
  7362. [0130 18:07:52 @multigpu.py:323] [p0574]  step: count(1578), step_time 1443.19, mean_step_time 1433.99, it/s 0.7
  7363. [0130 18:07:52 @multigpu.py:323] [p0576]  step: count(1587), step_time 1410.31, mean_step_time 1442.58, it/s 0.69
  7364. [0130 18:07:54 @multigpu.py:323] [p0115]  step: count(1564), step_time 1425.23, mean_step_time 1429.14, it/s 0.7
  7365. [0130 18:07:54 @multigpu.py:323] [p0576]  step: count(1588), step_time 1406.86, mean_step_time 1443.36, it/s 0.69
  7366. [0130 18:07:54 @multigpu.py:323] [p0574]  step: count(1579), step_time 1450.85, mean_step_time 1435.38, it/s 0.7
  7367. [0130 18:07:55 @multigpu.py:323] [p0576]  step: count(1589), step_time 1412.81, mean_step_time 1444.67, it/s 0.69
  7368. [0130 18:07:55 @multigpu.py:323] [p0115]  step: count(1565), step_time 1504.99, mean_step_time 1433.67, it/s 0.7
  7369. [0130 18:07:55 @multigpu.py:323] [p0574]  step: count(1580), step_time 1418.11, mean_step_time 1436.35, it/s 0.7
  7370. [0130 18:07:57 @multigpu.py:323] [p0115]  step: count(1566), step_time 1413.67, mean_step_time 1431.1, it/s 0.7
  7371. [0130 18:07:57 @multigpu.py:323] [p0576]  step: count(1590), step_time 1489.01, mean_step_time 1446.91, it/s 0.69
  7372. [0130 18:07:57 @multigpu.py:323] [p0574]  step: count(1581), step_time 1467.07, mean_step_time 1440.17, it/s 0.69
  7373. [0130 18:07:58 @multigpu.py:323] [p0115]  step: count(1567), step_time 1436.95, mean_step_time 1434.15, it/s 0.7
  7374. [0130 18:07:58 @multigpu.py:323] [p0576]  step: count(1591), step_time 1394.18, mean_step_time 1446.96, it/s 0.69
  7375. [0130 18:07:58 @multigpu.py:323] [p0574]  step: count(1582), step_time 1429.64, mean_step_time 1438.55, it/s 0.7
  7376. sending to address tcp://p0112:61216
  7377. ##### Sending to neptune:  online_score :  0.632678529951 , 2.6 #####
  7378. [u'online', 2.6]
    <