API tools faq
paste
Guest User

DFlash log

a guest
Jun 8th, 2026
57
0
Never
Add comment
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 44.81 KB | None | 0 0
raw download clone embed print report
  1. You are a helpful assistant<|im_end|>
  2. <|im_start|>user
  3. Hello<|im_end|>
  4. <|im_start|>assistant
  5. Hi there<|im_end|>
  6. <|im_start|>user
  7. How are you?<|im_end|>
  8. <|im_start|>assistant
  9. <think>
  10. '
  11. 0.12.898.672 I srv init: init: chat template, thinking = 1
  12. 0.12.898.714 I srv llama_server: model loaded
  13. 0.12.898.720 I srv llama_server: server is listening on http://127.0.0.1:8080
  14. 0.12.898.743 I srv update_slots: all slots are idle
  15. 1.02.438.922 I srv params_from_: Chat format: peg-native
  16. 1.02.473.882 I slot get_availabl: id 0 | task -1 | selected slot by LRU, t_last = -1
  17. 1.02.473.889 I slot get_availabl: id 0 | task -1 | adaptive dm: reset state for LRU slot selection
  18. 1.02.745.734 I srv recurrent_sh: shrunk recurrent state to 1 cells for prompt cache (before prompt cache save/load, removed 1 backup cells)
  19. 1.02.745.741 I srv get_availabl: updating prompt cache
  20. 1.02.745.747 I srv load: - looking for better prompt, base f_keep = -1.000, sim = 0.000
  21. 1.02.745.754 I srv update: - cache state: 0 prompts, 0.000 MiB (limits: 8192.000 MiB, 150016 tokens, 8589934592 est)
  22. 1.02.901.000 I srv recurrent_ex: expanded recurrent state to 2 cells after prompt cache (after prompt cache save/load)
  23. 1.02.901.007 I srv get_availabl: prompt cache update took 155.26 ms
  24. 1.02.901.532 I reasoning-budget: activated, budget=2147483647 tokens
  25. 1.02.901.556 I slot launch_slot_: id 0 | task 0 | processing task, is_child = 0
  26. 1.07.015.650 I slot print_timing: id 0 | task 0 | prompt processing, n_tokens = 6144, progress = 0.33, t = 4.11 s / 1493.48 tokens per second
  27. 1.08.290.208 I slot print_timing: id 0 | task 0 | prompt processing, n_tokens = 8192, progress = 0.44, t = 5.39 s / 1520.31 tokens per second
  28. 1.09.591.642 I slot print_timing: id 0 | task 0 | prompt processing, n_tokens = 10240, progress = 0.55, t = 6.69 s / 1530.70 tokens per second
  29. 1.10.913.418 I slot print_timing: id 0 | task 0 | prompt processing, n_tokens = 12288, progress = 0.66, t = 8.01 s / 1533.80 tokens per second
  30. 1.12.264.644 I slot print_timing: id 0 | task 0 | prompt processing, n_tokens = 14336, progress = 0.77, t = 9.36 s / 1531.19 tokens per second
  31. 1.13.650.993 I slot print_timing: id 0 | task 0 | prompt processing, n_tokens = 16384, progress = 0.88, t = 10.75 s / 1524.24 tokens per second
  32. 1.14.694.725 I dflash: drafter K/V projection cache enabled (1024-token window)
  33. 1.14.695.909 I slot print_timing: id 0 | task 0 | prompt processing, n_tokens = 17754, progress = 0.95, t = 11.79 s / 1505.36 tokens per second
  34. 1.14.762.550 I slot create_check: id 0 | task 0 | created context checkpoint 1 of 32 (pos_min = 17753, pos_max = 17753, n_tokens = 17754, size = 165.349 MiB)
  35. 1.15.045.068 I slot print_timing: id 0 | task 0 | prompt processing, n_tokens = 18101, progress = 0.97, t = 12.14 s / 1490.66 tokens per second
  36. 1.15.426.347 I slot print_timing: id 0 | task 0 | prompt processing, n_tokens = 18613, progress = 1.00, t = 12.52 s / 1486.16 tokens per second
  37. 1.15.634.581 I slot operator (): id 0 | task 0 | adaptive dm profit: cur=0 recommended=4 score=24.8 action=apply
  38. 1.15.831.630 I slot operator (): id 0 | task 0 | adaptive dm profit: cur=4 recommended=8 score=24.8 action=apply
  39. 1.16.027.807 I slot operator (): id 0 | task 0 | adaptive dm profit: cur=8 recommended=15 score=24.8 action=apply
  40. 1.16.441.851 I reasoning-budget: deactivated (natural end)
  41. 1.16.445.262 W slot process_toke: id 0 | task 0 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=41 reasoning_tokens=39 visible_tokens=2
  42. 1.17.700.534 I slot print_timing: id 0 | task 0 | prompt eval time = 12611.26 ms / 18617 tokens ( 0.68 ms per token, 1476.22 tokens per second)
  43. 1.17.700.540 I slot print_timing: id 0 | task 0 | eval time = 2186.27 ms / 84 tokens ( 26.03 ms per token, 38.42 tokens per second)
  44. 1.17.700.542 I slot print_timing: id 0 | task 0 | total time = 14797.53 ms / 18701 tokens
  45. 1.17.700.543 I slot print_timing: id 0 | task 0 | graphs reused = 28
  46. 1.17.700.544 I slot print_timing: id 0 | task 0 | draft acceptance = 0.10375 ( 47 accepted / 453 generated)
  47. 1.17.700.546 I slot print_timing: id 0 | task 0 | adaptive dm: fringe=0.00 n_max=15
  48. 1.17.700.573 I statistics dflash: #calls(b,g,a) = 1 34 34, #gen drafts = 34, #acc drafts = 23, #gen tokens = 453, #acc tokens = 47, dur(b,g,a) = 0.001, 290.493, 0.012 ms
  49. 1.17.701.658 I slot release: id 0 | task 0 | stop processing: n_tokens = 18701, truncated = 0
  50. 1.17.701.673 I srv update_slots: all slots are idle
  51. 1.18.045.997 I srv params_from_: Chat format: peg-native
  52. 1.18.075.001 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.890 (> 0.100 thold), f_keep = 0.998
  53. 1.18.075.015 I slot get_availabl: id 0 | task -1 | adaptive dm: reset state for prompt change (sim=0.890, keep=0.998, preserve_policy=0, n_max=15)
  54. 1.18.075.496 I reasoning-budget: activated, budget=2147483647 tokens
  55. 1.18.075.691 I slot launch_slot_: id 0 | task 50 | processing task, is_child = 0
  56. 1.18.075.743 I slot update_slots: id 0 | task 50 | Checking checkpoint with [17753, 17753] against 18658...
  57. 1.18.129.036 W slot update_slots: id 0 | task 50 | restored context checkpoint (pos_min = 17753, pos_max = 17753, n_tokens = 17754, n_past = 17754, size = 165.349 MiB)
  58. 1.18.131.418 W find_slot: non-consecutive token position 18265 after 17754 for sequence 0 with 512 new tokens
  59. 1.18.132.201 W find_slot: non-consecutive token position 18265 after 17754 for sequence 0 with 512 new tokens
  60. 1.20.171.131 I slot create_check: id 0 | task 50 | created context checkpoint 2 of 32 (pos_min = 20442, pos_max = 20442, n_tokens = 20443, size = 214.958 MiB)
  61. 1.22.680.108 I slot print_timing: id 0 | task 50 | n_decoded = 102, tg = 49.36 t/s
  62. 1.25.712.901 I slot print_timing: id 0 | task 50 | n_decoded = 229, tg = 44.90 t/s
  63. 1.28.746.439 I slot print_timing: id 0 | task 50 | n_decoded = 350, tg = 43.04 t/s
  64. 1.29.723.324 I reasoning-budget: deactivated (natural end)
  65. 1.31.782.329 I slot print_timing: id 0 | task 50 | n_decoded = 499, tg = 44.68 t/s
  66. 1.33.924.195 W slot process_toke: id 0 | task 50 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=599 reasoning_tokens=387 visible_tokens=212
  67. 1.34.826.341 I slot print_timing: id 0 | task 50 | n_decoded = 628, tg = 44.18 t/s
  68. 1.37.417.093 I slot print_timing: id 0 | task 50 | prompt eval time = 2535.69 ms / 3205 tokens ( 0.79 ms per token, 1263.96 tokens per second)
  69. 1.37.417.100 I slot print_timing: id 0 | task 50 | eval time = 16803.84 ms / 746 tokens ( 22.53 ms per token, 44.39 tokens per second)
  70. 1.37.417.101 I slot print_timing: id 0 | task 50 | total time = 19339.52 ms / 3951 tokens
  71. 1.37.417.102 I slot print_timing: id 0 | task 50 | graphs reused = 286
  72. 1.37.417.104 I slot print_timing: id 0 | task 50 | draft acceptance = 0.09421 ( 433 accepted / 4596 generated)
  73. 1.37.417.106 I slot print_timing: id 0 | task 50 | adaptive dm: fringe=0.00 n_max=15
  74. 1.37.417.133 I statistics dflash: #calls(b,g,a) = 2 348 348, #gen drafts = 348, #acc drafts = 216, #gen tokens = 5049, #acc tokens = 480, dur(b,g,a) = 0.002, 2666.150, 0.110 ms
  75. 1.37.418.521 I slot release: id 0 | task 50 | stop processing: n_tokens = 21706, truncated = 0
  76. 1.37.418.663 I srv update_slots: all slots are idle
  77. 1.38.048.914 I srv params_from_: Chat format: peg-native
  78. 1.38.082.467 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.962 (> 0.100 thold), f_keep = 1.000
  79. 1.38.082.475 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.962, keep=1.000, n_max=15)
  80. 1.38.083.060 I reasoning-budget: activated, budget=2147483647 tokens
  81. 1.38.083.255 I slot launch_slot_: id 0 | task 369 | processing task, is_child = 0
  82. 1.38.169.102 I slot create_check: id 0 | task 369 | created context checkpoint 3 of 32 (pos_min = 21702, pos_max = 21702, n_tokens = 21706, size = 249.626 MiB)
  83. 1.39.298.961 I reasoning-budget: deactivated (natural end)
  84. 1.39.349.525 W slot process_toke: id 0 | task 369 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=26 reasoning_tokens=24 visible_tokens=2
  85. 1.40.983.023 I slot print_timing: id 0 | task 369 | n_decoded = 102, tg = 48.47 t/s
  86. 1.42.321.422 I slot print_timing: id 0 | task 369 | prompt eval time = 793.09 ms / 850 tokens ( 0.93 ms per token, 1071.76 tokens per second)
  87. 1.42.321.432 I slot print_timing: id 0 | task 369 | eval time = 3443.36 ms / 196 tokens ( 17.57 ms per token, 56.92 tokens per second)
  88. 1.42.321.433 I slot print_timing: id 0 | task 369 | total time = 4236.45 ms / 1046 tokens
  89. 1.42.321.434 I slot print_timing: id 0 | task 369 | graphs reused = 335
  90. 1.42.321.436 I slot print_timing: id 0 | task 369 | draft acceptance = 0.15525 ( 136 accepted / 876 generated)
  91. 1.42.321.438 I slot print_timing: id 0 | task 369 | adaptive dm: fringe=0.00 n_max=15
  92. 1.42.321.461 I statistics dflash: #calls(b,g,a) = 3 408 408, #gen drafts = 408, #acc drafts = 256, #gen tokens = 5925, #acc tokens = 616, dur(b,g,a) = 0.003, 3116.794, 0.132 ms
  93. 1.42.323.063 I slot release: id 0 | task 369 | stop processing: n_tokens = 22752, truncated = 0
  94. 1.42.323.303 I srv update_slots: all slots are idle
  95. 1.42.949.258 I srv params_from_: Chat format: peg-native
  96. 1.42.983.596 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.963 (> 0.100 thold), f_keep = 0.993
  97. 1.42.983.601 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.963, keep=0.993, n_max=15)
  98. 1.42.984.082 I reasoning-budget: activated, budget=2147483647 tokens
  99. 1.42.984.260 I slot launch_slot_: id 0 | task 433 | processing task, is_child = 0
  100. 1.42.984.322 I slot update_slots: id 0 | task 433 | Checking checkpoint with [21702, 21702] against 22582...
  101. 1.43.079.248 W slot update_slots: id 0 | task 433 | restored context checkpoint (pos_min = 21702, pos_max = 21702, n_tokens = 21706, n_past = 21703, size = 249.626 MiB)
  102. 1.43.083.951 W find_slot: non-consecutive token position 22214 after 21703 for sequence 0 with 512 new tokens
  103. 1.43.084.594 W find_slot: non-consecutive token position 22214 after 21703 for sequence 0 with 512 new tokens
  104. 1.43.999.205 I slot create_check: id 0 | task 433 | created context checkpoint 4 of 32 (pos_min = 22752, pos_max = 22752, n_tokens = 22753, size = 249.626 MiB)
  105. 1.46.700.541 I slot print_timing: id 0 | task 433 | n_decoded = 101, tg = 48.91 t/s
  106. 1.49.725.566 I slot print_timing: id 0 | task 433 | n_decoded = 220, tg = 43.21 t/s
  107. 1.50.363.756 I slot operator (): id 0 | task 433 | adaptive dm profit: cur=15 recommended=13 score=48.7 action=apply
  108. 1.50.688.706 I slot operator (): id 0 | task 433 | adaptive dm profit: cur=13 recommended=11 score=48.7 action=apply
  109. 1.51.128.032 I slot operator (): id 0 | task 433 | adaptive dm profit: cur=11 recommended=10 score=48.7 action=apply
  110. 1.51.446.985 I slot operator (): id 0 | task 433 | adaptive dm profit: cur=10 recommended=14 score=46.9 action=apply
  111. 1.51.900.965 I slot operator (): id 0 | task 433 | adaptive dm profit: cur=14 recommended=15 score=46.9 action=apply
  112. 1.52.736.204 I slot print_timing: id 0 | task 433 | n_decoded = 321, tg = 39.63 t/s
  113. 1.53.097.979 I reasoning-budget: deactivated (natural end)
  114. 1.53.560.854 W slot process_toke: id 0 | task 433 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=350 reasoning_tokens=332 visible_tokens=18
  115. 1.55.750.858 I slot print_timing: id 0 | task 433 | n_decoded = 440, tg = 39.58 t/s
  116. 1.56.260.415 I slot print_timing: id 0 | task 433 | prompt eval time = 1648.77 ms / 1750 tokens ( 0.94 ms per token, 1061.40 tokens per second)
  117. 1.56.260.426 I slot print_timing: id 0 | task 433 | eval time = 11626.05 ms / 470 tokens ( 24.74 ms per token, 40.43 tokens per second)
  118. 1.56.260.428 I slot print_timing: id 0 | task 433 | total time = 13274.81 ms / 2220 tokens
  119. 1.56.260.429 I slot print_timing: id 0 | task 433 | graphs reused = 512
  120. 1.56.260.430 I slot print_timing: id 0 | task 433 | draft acceptance = 0.08585 ( 259 accepted / 3017 generated)
  121. 1.56.260.432 I slot print_timing: id 0 | task 433 | adaptive dm: fringe=0.00 n_max=15
  122. 1.56.260.452 I statistics dflash: #calls(b,g,a) = 4 619 619, #gen drafts = 619, #acc drafts = 392, #gen tokens = 8942, #acc tokens = 875, dur(b,g,a) = 0.004, 4704.318, 0.218 ms
  123. 1.56.261.717 I slot release: id 0 | task 433 | stop processing: n_tokens = 23923, truncated = 0
  124. 1.56.261.831 I srv update_slots: all slots are idle
  125. 1.56.569.617 I srv params_from_: Chat format: peg-native
  126. 1.56.609.319 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.988 (> 0.100 thold), f_keep = 0.995
  127. 1.56.609.327 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.988, keep=0.995, n_max=15)
  128. 1.56.609.799 I reasoning-budget: activated, budget=2147483647 tokens
  129. 1.56.609.971 I slot launch_slot_: id 0 | task 649 | processing task, is_child = 0
  130. 1.56.610.052 I slot update_slots: id 0 | task 649 | Checking checkpoint with [22752, 22752] against 23803...
  131. 1.56.690.247 W slot update_slots: id 0 | task 649 | restored context checkpoint (pos_min = 22752, pos_max = 22752, n_tokens = 22753, n_past = 22753, size = 249.626 MiB)
  132. 1.56.695.039 W find_slot: non-consecutive token position 23264 after 22753 for sequence 0 with 512 new tokens
  133. 1.56.696.040 W find_slot: non-consecutive token position 23264 after 22753 for sequence 0 with 512 new tokens
  134. 1.57.741.073 I slot create_check: id 0 | task 649 | created context checkpoint 5 of 32 (pos_min = 23923, pos_max = 23923, n_tokens = 23924, size = 249.626 MiB)
  135. 1.58.230.952 I reasoning-budget: deactivated (natural end)
  136. 1.58.232.170 W slot process_toke: id 0 | task 649 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=13 reasoning_tokens=11 visible_tokens=2
  137. 1.59.531.939 I slot print_timing: id 0 | task 649 | n_decoded = 105, tg = 66.63 t/s
  138. 1.59.973.247 I slot print_timing: id 0 | task 649 | prompt eval time = 1345.00 ms / 1340 tokens ( 1.00 ms per token, 996.28 tokens per second)
  139. 1.59.973.255 I slot print_timing: id 0 | task 649 | eval time = 2017.16 ms / 134 tokens ( 15.05 ms per token, 66.43 tokens per second)
  140. 1.59.973.261 I slot print_timing: id 0 | task 649 | total time = 3362.16 ms / 1474 tokens
  141. 1.59.973.264 I slot print_timing: id 0 | task 649 | graphs reused = 543
  142. 1.59.973.266 I slot print_timing: id 0 | task 649 | draft acceptance = 0.20808 ( 103 accepted / 495 generated)
  143. 1.59.973.268 I slot print_timing: id 0 | task 649 | adaptive dm: fringe=0.00 n_max=15
  144. 1.59.973.296 I statistics dflash: #calls(b,g,a) = 5 652 652, #gen drafts = 652, #acc drafts = 418, #gen tokens = 9437, #acc tokens = 978, dur(b,g,a) = 0.005, 4949.271, 0.227 ms
  145. 1.59.974.628 I slot release: id 0 | task 649 | stop processing: n_tokens = 24229, truncated = 0
  146. 1.59.974.749 I srv update_slots: all slots are idle
  147. 2.00.293.682 I srv params_from_: Chat format: peg-native
  148. 2.00.331.927 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.988 (> 0.100 thold), f_keep = 0.995
  149. 2.00.331.933 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.988, keep=0.995, n_max=15)
  150. 2.00.332.410 I reasoning-budget: activated, budget=2147483647 tokens
  151. 2.00.332.575 I slot launch_slot_: id 0 | task 687 | processing task, is_child = 0
  152. 2.00.332.650 I slot update_slots: id 0 | task 687 | Checking checkpoint with [23923, 23923] against 24106...
  153. 2.00.413.502 W slot update_slots: id 0 | task 687 | restored context checkpoint (pos_min = 23923, pos_max = 23923, n_tokens = 23924, n_past = 23924, size = 249.626 MiB)
  154. 2.00.418.255 W find_slot: non-consecutive token position 24227 after 23926 for sequence 0 with 304 new tokens
  155. 2.00.418.575 W find_slot: non-consecutive token position 24227 after 23926 for sequence 0 with 304 new tokens
  156. 2.01.820.575 I reasoning-budget: deactivated (natural end)
  157. 2.01.821.774 W slot process_toke: id 0 | task 687 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=40 reasoning_tokens=38 visible_tokens=2
  158. 2.03.621.847 I slot print_timing: id 0 | task 687 | n_decoded = 101, tg = 37.27 t/s
  159. 2.06.646.854 I slot print_timing: id 0 | task 687 | n_decoded = 310, tg = 54.06 t/s
  160. 2.09.668.611 I slot print_timing: id 0 | task 687 | n_decoded = 559, tg = 63.84 t/s
  161. 2.12.705.269 I slot print_timing: id 0 | task 687 | n_decoded = 918, tg = 77.84 t/s
  162. 2.15.745.648 I slot print_timing: id 0 | task 687 | n_decoded = 1259, tg = 84.88 t/s
  163. 2.18.675.350 I slot print_timing: id 0 | task 687 | prompt eval time = 578.36 ms / 474 tokens ( 1.22 ms per token, 819.56 tokens per second)
  164. 2.18.675.356 I slot print_timing: id 0 | task 687 | eval time = 17762.74 ms / 1517 tokens ( 11.71 ms per token, 85.40 tokens per second)
  165. 2.18.675.357 I slot print_timing: id 0 | task 687 | total time = 18341.10 ms / 1991 tokens
  166. 2.18.675.357 I slot print_timing: id 0 | task 687 | graphs reused = 877
  167. 2.18.675.359 I slot print_timing: id 0 | task 687 | draft acceptance = 0.22904 ( 1175 accepted / 5130 generated)
  168. 2.18.675.360 I slot print_timing: id 0 | task 687 | adaptive dm: fringe=0.00 n_max=15
  169. 2.18.675.386 I statistics dflash: #calls(b,g,a) = 6 994 994, #gen drafts = 994, #acc drafts = 681, #gen tokens = 14567, #acc tokens = 2153, dur(b,g,a) = 0.006, 7465.196, 0.323 ms
  170. 2.18.676.478 I slot release: id 0 | task 687 | stop processing: n_tokens = 25915, truncated = 0
  171. 2.18.676.590 I srv update_slots: all slots are idle
  172. 2.19.164.042 I srv params_from_: Chat format: peg-native
  173. 2.19.208.248 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.941 (> 0.100 thold), f_keep = 0.943
  174. 2.19.208.253 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.941, keep=0.943, n_max=15)
  175. 2.19.208.921 I reasoning-budget: activated, budget=2147483647 tokens
  176. 2.19.209.117 I slot launch_slot_: id 0 | task 1033 | processing task, is_child = 0
  177. 2.19.209.211 I slot update_slots: id 0 | task 1033 | Checking checkpoint with [23923, 23923] against 24438...
  178. 2.19.295.507 W slot update_slots: id 0 | task 1033 | restored context checkpoint (pos_min = 23923, pos_max = 23923, n_tokens = 23924, n_past = 23924, size = 249.626 MiB)
  179. 2.19.300.391 W find_slot: non-consecutive token position 24435 after 23924 for sequence 0 with 512 new tokens
  180. 2.19.301.311 W find_slot: non-consecutive token position 24435 after 23924 for sequence 0 with 512 new tokens
  181. 2.20.947.352 I slot create_check: id 0 | task 1033 | created context checkpoint 6 of 32 (pos_min = 25915, pos_max = 25915, n_tokens = 25916, size = 249.626 MiB)
  182. 2.21.496.367 I reasoning-budget: deactivated (natural end)
  183. 2.21.498.503 W slot process_toke: id 0 | task 1033 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=20 reasoning_tokens=18 visible_tokens=2
  184. 2.23.551.318 I slot print_timing: id 0 | task 1033 | n_decoded = 107, tg = 43.16 t/s
  185. 2.24.414.343 I slot print_timing: id 0 | task 1033 | prompt eval time = 1862.26 ms / 2034 tokens ( 0.92 ms per token, 1092.22 tokens per second)
  186. 2.24.414.350 I slot print_timing: id 0 | task 1033 | eval time = 3341.76 ms / 141 tokens ( 23.70 ms per token, 42.19 tokens per second)
  187. 2.24.414.352 I slot print_timing: id 0 | task 1033 | total time = 5204.02 ms / 2175 tokens
  188. 2.24.414.353 I slot print_timing: id 0 | task 1033 | graphs reused = 933
  189. 2.24.414.355 I slot print_timing: id 0 | task 1033 | draft acceptance = 0.09492 ( 84 accepted / 885 generated)
  190. 2.24.414.356 I slot print_timing: id 0 | task 1033 | adaptive dm: fringe=0.00 n_max=15
  191. 2.24.414.379 I statistics dflash: #calls(b,g,a) = 7 1053 1053, #gen drafts = 1053, #acc drafts = 718, #gen tokens = 15452, #acc tokens = 2237, dur(b,g,a) = 0.007, 7895.826, 0.345 ms
  192. 2.24.415.758 I slot release: id 0 | task 1033 | stop processing: n_tokens = 26101, truncated = 0
  193. 2.24.415.868 I srv update_slots: all slots are idle
  194. 2.24.765.307 I srv params_from_: Chat format: peg-native
  195. 2.24.803.192 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.989 (> 0.100 thold), f_keep = 0.995
  196. 2.24.803.198 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.989, keep=0.995, n_max=15)
  197. 2.24.803.675 I reasoning-budget: activated, budget=2147483647 tokens
  198. 2.24.803.833 I slot launch_slot_: id 0 | task 1097 | processing task, is_child = 0
  199. 2.24.803.908 I slot update_slots: id 0 | task 1097 | Checking checkpoint with [25915, 25915] against 25978...
  200. 2.24.884.103 W slot update_slots: id 0 | task 1097 | restored context checkpoint (pos_min = 25915, pos_max = 25915, n_tokens = 25916, n_past = 25916, size = 249.626 MiB)
  201. 2.24.888.649 W find_slot: non-consecutive token position 26099 after 25918 for sequence 0 with 184 new tokens
  202. 2.24.888.831 W find_slot: non-consecutive token position 26099 after 25918 for sequence 0 with 184 new tokens
  203. 2.25.575.787 I reasoning-budget: deactivated (natural end)
  204. 2.25.874.954 W slot process_toke: id 0 | task 1097 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=31 reasoning_tokens=16 visible_tokens=15
  205. 2.25.923.412 I slot operator (): id 0 | task 1097 | adaptive dm profit: cur=0 recommended=15 score=60.1 action=apply
  206. 2.27.515.313 I slot print_timing: id 0 | task 1097 | n_decoded = 102, tg = 45.93 t/s
  207. 2.30.524.660 I slot print_timing: id 0 | task 1097 | n_decoded = 290, tg = 55.46 t/s
  208. 2.33.538.840 I slot print_timing: id 0 | task 1097 | n_decoded = 536, tg = 65.03 t/s
  209. 2.36.564.723 I slot print_timing: id 0 | task 1097 | n_decoded = 796, tg = 70.63 t/s
  210. 2.39.567.122 I slot print_timing: id 0 | task 1097 | n_decoded = 987, tg = 69.16 t/s
  211. 2.42.580.883 I slot print_timing: id 0 | task 1097 | n_decoded = 1160, tg = 67.11 t/s
  212. 2.43.390.985 I slot print_timing: id 0 | task 1097 | prompt eval time = 489.40 ms / 354 tokens ( 1.38 ms per token, 723.34 tokens per second)
  213. 2.43.390.992 I slot print_timing: id 0 | task 1097 | eval time = 18094.44 ms / 1225 tokens ( 14.77 ms per token, 67.70 tokens per second)
  214. 2.43.390.994 I slot print_timing: id 0 | task 1097 | total time = 18583.84 ms / 1579 tokens
  215. 2.43.390.994 I slot print_timing: id 0 | task 1097 | graphs reused = 1268
  216. 2.43.390.996 I slot print_timing: id 0 | task 1097 | draft acceptance = 0.17173 ( 881 accepted / 5130 generated)
  217. 2.43.390.997 I slot print_timing: id 0 | task 1097 | adaptive dm: fringe=0.00 n_max=15
  218. 2.43.391.019 I statistics dflash: #calls(b,g,a) = 8 1395 1395, #gen drafts = 1395, #acc drafts = 984, #gen tokens = 20582, #acc tokens = 3118, dur(b,g,a) = 0.008, 10435.572, 0.447 ms
  219. 2.43.392.459 I slot release: id 0 | task 1097 | stop processing: n_tokens = 27494, truncated = 0
  220. 2.43.392.663 I srv update_slots: all slots are idle
  221. 2.44.367.140 I srv params_from_: Chat format: peg-native
  222. 2.44.408.075 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.998 (> 0.100 thold), f_keep = 1.000
  223. 2.44.408.081 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.998, keep=1.000, n_max=15)
  224. 2.44.408.562 I reasoning-budget: activated, budget=2147483647 tokens
  225. 2.44.408.735 I slot launch_slot_: id 0 | task 1444 | processing task, is_child = 0
  226. 2.44.533.266 I slot create_check: id 0 | task 1444 | created context checkpoint 7 of 32 (pos_min = 27495, pos_max = 27495, n_tokens = 27496, size = 149.821 MiB)
  227. 2.46.906.330 I slot print_timing: id 0 | task 1444 | n_decoded = 104, tg = 46.36 t/s
  228. 2.47.473.097 I reasoning-budget: deactivated (natural end)
  229. 2.48.186.015 W slot process_toke: id 0 | task 1444 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=160 reasoning_tokens=130 visible_tokens=30
  230. 2.49.911.118 I slot print_timing: id 0 | task 1444 | n_decoded = 218, tg = 41.54 t/s
  231. 2.50.177.359 I slot print_timing: id 0 | task 1444 | prompt eval time = 251.65 ms / 42 tokens ( 5.99 ms per token, 166.90 tokens per second)
  232. 2.50.177.366 I slot print_timing: id 0 | task 1444 | eval time = 5514.76 ms / 230 tokens ( 23.98 ms per token, 41.71 tokens per second)
  233. 2.50.177.368 I slot print_timing: id 0 | task 1444 | total time = 5766.41 ms / 272 tokens
  234. 2.50.177.369 I slot print_timing: id 0 | task 1444 | graphs reused = 1370
  235. 2.50.177.370 I slot print_timing: id 0 | task 1444 | draft acceptance = 0.07873 ( 124 accepted / 1575 generated)
  236. 2.50.177.372 I slot print_timing: id 0 | task 1444 | adaptive dm: fringe=0.00 n_max=15
  237. 2.50.177.393 I statistics dflash: #calls(b,g,a) = 9 1500 1500, #gen drafts = 1500, #acc drafts = 1041, #gen tokens = 22157, #acc tokens = 3242, dur(b,g,a) = 0.008, 11146.553, 0.478 ms
  238. 2.50.178.826 I slot release: id 0 | task 1444 | stop processing: n_tokens = 27765, truncated = 0
  239. 2.50.178.967 I srv update_slots: all slots are idle
  240. 2.50.506.885 I srv params_from_: Chat format: peg-native
  241. 2.50.550.300 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.995 (> 0.100 thold), f_keep = 0.998
  242. 2.50.550.306 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.995, keep=0.998, n_max=15)
  243. 2.50.550.803 I reasoning-budget: activated, budget=2147483647 tokens
  244. 2.50.550.996 I slot launch_slot_: id 0 | task 1553 | processing task, is_child = 0
  245. 2.50.551.087 I slot update_slots: id 0 | task 1553 | Checking checkpoint with [27495, 27495] against 27704...
  246. 2.50.598.529 W slot update_slots: id 0 | task 1553 | restored context checkpoint (pos_min = 27495, pos_max = 27495, n_tokens = 27496, n_past = 27496, size = 149.821 MiB)
  247. 2.50.600.337 W find_slot: non-consecutive token position 27765 after 27496 for sequence 0 with 270 new tokens
  248. 2.50.600.605 W find_slot: non-consecutive token position 27765 after 27496 for sequence 0 with 270 new tokens
  249. 2.51.379.694 I reasoning-budget: deactivated (natural end)
  250. 2.51.382.523 W slot process_toke: id 0 | task 1553 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=27 reasoning_tokens=25 visible_tokens=2
  251. 2.52.941.281 I slot print_timing: id 0 | task 1553 | n_decoded = 123, tg = 63.19 t/s
  252. 2.53.592.749 I slot print_timing: id 0 | task 1553 | prompt eval time = 441.79 ms / 360 tokens ( 1.23 ms per token, 814.88 tokens per second)
  253. 2.53.592.756 I slot print_timing: id 0 | task 1553 | eval time = 2598.02 ms / 199 tokens ( 13.06 ms per token, 76.60 tokens per second)
  254. 2.53.592.757 I slot print_timing: id 0 | task 1553 | total time = 3039.81 ms / 559 tokens
  255. 2.53.592.759 I slot print_timing: id 0 | task 1553 | graphs reused = 1414
  256. 2.53.592.760 I slot print_timing: id 0 | task 1553 | draft acceptance = 0.21560 ( 152 accepted / 705 generated)
  257. 2.53.592.761 I slot print_timing: id 0 | task 1553 | adaptive dm: fringe=0.00 n_max=15
  258. 2.53.592.835 I statistics dflash: #calls(b,g,a) = 10 1547 1547, #gen drafts = 1547, #acc drafts = 1079, #gen tokens = 22862, #acc tokens = 3394, dur(b,g,a) = 0.009, 11486.891, 0.495 ms
  259. 2.53.594.550 I slot release: id 0 | task 1553 | stop processing: n_tokens = 28055, truncated = 0
  260. 2.53.594.662 I srv update_slots: all slots are idle
  261. 2.53.983.175 I srv params_from_: Chat format: peg-native
  262. 2.54.042.824 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.990 (> 0.100 thold), f_keep = 0.994
  263. 2.54.042.830 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.990, keep=0.994, n_max=15)
  264. 2.54.043.369 I reasoning-budget: activated, budget=2147483647 tokens
  265. 2.54.043.542 I slot launch_slot_: id 0 | task 1604 | processing task, is_child = 0
  266. 2.54.043.611 I slot update_slots: id 0 | task 1604 | Checking checkpoint with [27495, 27495] against 27883...
  267. 2.54.095.022 W slot update_slots: id 0 | task 1604 | restored context checkpoint (pos_min = 27495, pos_max = 27495, n_tokens = 27496, n_past = 27496, size = 149.821 MiB)
  268. 2.54.096.890 W find_slot: non-consecutive token position 27642 after 27496 for sequence 0 with 147 new tokens
  269. 2.54.097.059 W find_slot: non-consecutive token position 27642 after 27496 for sequence 0 with 147 new tokens
  270. 2.55.110.709 I reasoning-budget: deactivated (natural end)
  271. 2.55.163.720 W slot process_toke: id 0 | task 1604 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=16 reasoning_tokens=14 visible_tokens=2
  272. 2.56.481.590 I slot print_timing: id 0 | task 1604 | n_decoded = 104, tg = 62.77 t/s
  273. 2.57.387.327 I slot print_timing: id 0 | task 1604 | prompt eval time = 778.28 ms / 663 tokens ( 1.17 ms per token, 851.87 tokens per second)
  274. 2.57.387.333 I slot print_timing: id 0 | task 1604 | eval time = 2564.45 ms / 188 tokens ( 13.64 ms per token, 73.31 tokens per second)
  275. 2.57.387.335 I slot print_timing: id 0 | task 1604 | total time = 3342.73 ms / 851 tokens
  276. 2.57.387.335 I slot print_timing: id 0 | task 1604 | graphs reused = 1457
  277. 2.57.387.338 I slot print_timing: id 0 | task 1604 | draft acceptance = 0.21185 ( 143 accepted / 675 generated)
  278. 2.57.387.339 I slot print_timing: id 0 | task 1604 | adaptive dm: fringe=0.00 n_max=15
  279. 2.57.387.365 I statistics dflash: #calls(b,g,a) = 11 1592 1592, #gen drafts = 1592, #acc drafts = 1113, #gen tokens = 23537, #acc tokens = 3537, dur(b,g,a) = 0.009, 11815.407, 0.507 ms
  280. 2.57.389.094 I slot release: id 0 | task 1604 | stop processing: n_tokens = 28347, truncated = 0
  281. 2.57.389.199 I srv update_slots: all slots are idle
  282. 2.57.811.137 I srv params_from_: Chat format: peg-native
  283. 2.57.864.908 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.990 (> 0.100 thold), f_keep = 0.994
  284. 2.57.864.916 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.990, keep=0.994, n_max=15)
  285. 2.57.865.605 I reasoning-budget: activated, budget=2147483647 tokens
  286. 2.57.865.961 I slot launch_slot_: id 0 | task 1654 | processing task, is_child = 0
  287. 2.57.866.083 I slot update_slots: id 0 | task 1654 | Checking checkpoint with [27495, 27495] against 28175...
  288. 2.57.919.850 W slot update_slots: id 0 | task 1654 | restored context checkpoint (pos_min = 27495, pos_max = 27495, n_tokens = 27496, n_past = 27496, size = 149.821 MiB)
  289. 2.57.922.303 W find_slot: non-consecutive token position 27946 after 27496 for sequence 0 with 451 new tokens
  290. 2.57.922.784 W find_slot: non-consecutive token position 27946 after 27496 for sequence 0 with 451 new tokens
  291. 2.59.090.745 I reasoning-budget: deactivated (natural end)
  292. 2.59.092.316 W slot process_toke: id 0 | task 1654 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=14 reasoning_tokens=12 visible_tokens=2
  293. 3.00.840.579 I slot print_timing: id 0 | task 1654 | n_decoded = 107, tg = 53.71 t/s
  294. 3.01.079.156 I slot print_timing: id 0 | task 1654 | prompt eval time = 980.16 ms / 967 tokens ( 1.01 ms per token, 986.57 tokens per second)
  295. 3.01.079.163 I slot print_timing: id 0 | task 1654 | eval time = 2231.99 ms / 113 tokens ( 19.75 ms per token, 50.63 tokens per second)
  296. 3.01.079.164 I slot print_timing: id 0 | task 1654 | total time = 3212.15 ms / 1080 tokens
  297. 3.01.079.165 I slot print_timing: id 0 | task 1654 | graphs reused = 1491
  298. 3.01.079.167 I slot print_timing: id 0 | task 1654 | draft acceptance = 0.14630 ( 79 accepted / 540 generated)
  299. 3.01.079.168 I slot print_timing: id 0 | task 1654 | adaptive dm: fringe=0.00 n_max=15
  300. 3.01.079.190 I statistics dflash: #calls(b,g,a) = 12 1628 1628, #gen drafts = 1628, #acc drafts = 1140, #gen tokens = 24077, #acc tokens = 3616, dur(b,g,a) = 0.009, 12080.854, 0.524 ms
  301. 3.01.080.557 I slot release: id 0 | task 1654 | stop processing: n_tokens = 28578, truncated = 0
  302. 3.01.080.671 I srv update_slots: all slots are idle
  303. 3.01.526.877 I srv params_from_: Chat format: peg-native
  304. 3.01.590.643 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.992 (> 0.100 thold), f_keep = 0.997
  305. 3.01.590.650 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.992, keep=0.997, n_max=15)
  306. 3.01.591.168 I reasoning-budget: activated, budget=2147483647 tokens
  307. 3.01.591.351 I slot launch_slot_: id 0 | task 1695 | processing task, is_child = 0
  308. 3.01.591.434 I slot update_slots: id 0 | task 1695 | Checking checkpoint with [27495, 27495] against 28477...
  309. 3.01.640.941 W slot update_slots: id 0 | task 1695 | restored context checkpoint (pos_min = 27495, pos_max = 27495, n_tokens = 27496, n_past = 27496, size = 149.821 MiB)
  310. 3.01.642.880 W find_slot: non-consecutive token position 28007 after 27498 for sequence 0 with 512 new tokens
  311. 3.01.643.436 W find_slot: non-consecutive token position 28007 after 27498 for sequence 0 with 512 new tokens
  312. 3.02.666.453 I slot create_check: id 0 | task 1695 | created context checkpoint 8 of 32 (pos_min = 28576, pos_max = 28576, n_tokens = 28577, size = 236.736 MiB)
  313. 3.03.256.065 I reasoning-budget: deactivated (natural end)
  314. 3.03.310.871 W slot process_toke: id 0 | task 1695 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=18 reasoning_tokens=16 visible_tokens=2
  315. 3.05.267.203 I slot print_timing: id 0 | task 1695 | n_decoded = 104, tg = 43.40 t/s
  316. 3.05.946.334 I slot print_timing: id 0 | task 1695 | prompt eval time = 1277.32 ms / 1215 tokens ( 1.05 ms per token, 951.21 tokens per second)
  317. 3.05.946.343 I slot print_timing: id 0 | task 1695 | eval time = 3076.59 ms / 139 tokens ( 22.13 ms per token, 45.18 tokens per second)
  318. 3.05.946.345 I slot print_timing: id 0 | task 1695 | total time = 4353.92 ms / 1354 tokens
  319. 3.05.946.347 I slot print_timing: id 0 | task 1695 | graphs reused = 1547
  320. 3.05.946.349 I slot print_timing: id 0 | task 1695 | draft acceptance = 0.09310 ( 81 accepted / 870 generated)
  321. 3.05.946.350 I slot print_timing: id 0 | task 1695 | adaptive dm: fringe=0.00 n_max=15
  322. 3.05.946.373 I statistics dflash: #calls(b,g,a) = 13 1686 1686, #gen drafts = 1686, #acc drafts = 1178, #gen tokens = 24947, #acc tokens = 3697, dur(b,g,a) = 0.009, 12507.594, 0.535 ms
  323. 3.05.947.792 I slot release: id 0 | task 1695 | stop processing: n_tokens = 28850, truncated = 0
  324. 3.05.947.913 I srv update_slots: all slots are idle
  325. 3.06.271.465 I srv params_from_: Chat format: peg-native
  326. 3.06.314.397 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.990 (> 0.100 thold), f_keep = 0.996
  327. 3.06.314.402 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.990, keep=0.996, n_max=15)
  328. 3.06.314.879 I reasoning-budget: activated, budget=2147483647 tokens
  329. 3.06.315.046 I slot launch_slot_: id 0 | task 1758 | processing task, is_child = 0
  330. 3.06.315.129 I slot update_slots: id 0 | task 1758 | Checking checkpoint with [28576, 28576] against 28729...
  331. 3.06.392.915 W slot update_slots: id 0 | task 1758 | restored context checkpoint (pos_min = 28576, pos_max = 28576, n_tokens = 28577, n_past = 28577, size = 236.736 MiB)
  332. 3.06.397.081 W find_slot: non-consecutive token position 28850 after 28577 for sequence 0 with 274 new tokens
  333. 3.06.397.342 W find_slot: non-consecutive token position 28850 after 28577 for sequence 0 with 274 new tokens
  334. 3.07.192.065 I reasoning-budget: deactivated (natural end)
  335. 3.07.244.169 W slot process_toke: id 0 | task 1758 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=19 reasoning_tokens=17 visible_tokens=2
  336. 3.08.997.709 I slot print_timing: id 0 | task 1758 | prompt eval time = 540.91 ms / 444 tokens ( 1.22 ms per token, 820.85 tokens per second)
  337. 3.08.997.716 I slot print_timing: id 0 | task 1758 | eval time = 2140.67 ms / 84 tokens ( 25.48 ms per token, 39.24 tokens per second)
  338. 3.08.997.717 I slot print_timing: id 0 | task 1758 | total time = 2681.58 ms / 528 tokens
  339. 3.08.997.718 I slot print_timing: id 0 | task 1758 | graphs reused = 1581
  340. 3.08.997.720 I slot print_timing: id 0 | task 1758 | draft acceptance = 0.09259 ( 50 accepted / 540 generated)
  341. 3.08.997.721 I slot print_timing: id 0 | task 1758 | adaptive dm: fringe=0.00 n_max=15
  342. 3.08.997.743 I statistics dflash: #calls(b,g,a) = 14 1722 1722, #gen drafts = 1722, #acc drafts = 1205, #gen tokens = 25487, #acc tokens = 3747, dur(b,g,a) = 0.010, 12769.958, 0.545 ms
  343. 3.08.999.193 I slot release: id 0 | task 1758 | stop processing: n_tokens = 29107, truncated = 0
  344. 3.08.999.323 I srv update_slots: all slots are idle
  345. 3.13.522.808 I srv params_from_: Chat format: peg-native
  346. 3.13.562.555 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.996 (> 0.100 thold), f_keep = 0.998
  347. 3.13.562.561 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.996, keep=0.998, n_max=15)
  348. 3.13.563.033 I reasoning-budget: activated, budget=2147483647 tokens
  349. 3.13.563.212 I slot launch_slot_: id 0 | task 1798 | processing task, is_child = 0
  350. 3.13.563.309 I slot update_slots: id 0 | task 1798 | Checking checkpoint with [28576, 28576] against 29040...
  351. 3.13.643.848 W slot update_slots: id 0 | task 1798 | restored context checkpoint (pos_min = 28576, pos_max = 28576, n_tokens = 28577, n_past = 28577, size = 236.736 MiB)
  352. 3.13.648.491 W find_slot: non-consecutive token position 28654 after 28579 for sequence 0 with 78 new tokens
  353. 3.13.648.596 W find_slot: non-consecutive token position 28654 after 28579 for sequence 0 with 78 new tokens
  354. 3.14.614.622 I reasoning-budget: deactivated (natural end)
  355. 3.14.666.561 W slot process_toke: id 0 | task 1798 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=16 reasoning_tokens=14 visible_tokens=2
  356. 3.15.528.971 I slot print_timing: id 0 | task 1798 | prompt eval time = 810.72 ms / 594 tokens ( 1.36 ms per token, 732.69 tokens per second)
  357. 3.15.528.979 I slot print_timing: id 0 | task 1798 | eval time = 1153.89 ms / 76 tokens ( 15.18 ms per token, 65.86 tokens per second)
  358. 3.15.528.980 I slot print_timing: id 0 | task 1798 | total time = 1964.60 ms / 670 tokens
  359. 3.15.528.982 I slot print_timing: id 0 | task 1798 | graphs reused = 1600
  360. 3.15.528.983 I slot print_timing: id 0 | task 1798 | draft acceptance = 0.17778 ( 56 accepted / 315 generated)
  361. 3.15.528.984 I slot print_timing: id 0 | task 1798 | adaptive dm: fringe=0.00 n_max=15
  362. 3.15.529.006 I statistics dflash: #calls(b,g,a) = 15 1743 1743, #gen drafts = 1743, #acc drafts = 1224, #gen tokens = 25802, #acc tokens = 3803, dur(b,g,a) = 0.011, 12923.533, 0.552 ms
  363. 3.15.530.260 I slot release: id 0 | task 1798 | stop processing: n_tokens = 29248, truncated = 0
  364. 3.15.530.441 I srv update_slots: all slots are idle
  365. 3.20.409.561 I srv params_from_: Chat format: peg-native
  366. 3.20.454.561 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.991 (> 0.100 thold), f_keep = 0.998
  367. 3.20.454.567 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.991, keep=0.998, n_max=15)
  368. 3.20.455.311 I reasoning-budget: activated, budget=2147483647 tokens
  369. 3.20.455.527 I slot launch_slot_: id 0 | task 1824 | processing task, is_child = 0
  370. 3.20.455.645 I slot update_slots: id 0 | task 1824 | Checking checkpoint with [28576, 28576] against 29187...
  371. 3.20.538.799 W slot update_slots: id 0 | task 1824 | restored context checkpoint (pos_min = 28576, pos_max = 28576, n_tokens = 28577, n_past = 28577, size = 236.736 MiB)
  372. 3.20.543.758 W find_slot: non-consecutive token position 28933 after 28578 for sequence 0 with 357 new tokens
  373. 3.20.544.109 W find_slot: non-consecutive token position 28933 after 28578 for sequence 0 with 357 new tokens
  374. 3.21.666.737 I reasoning-budget: deactivated (natural end)
  375. 3.21.718.740 W slot process_toke: id 0 | task 1824 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=16 reasoning_tokens=14 visible_tokens=2
  376. 3.23.349.205 I slot print_timing: id 0 | task 1824 | n_decoded = 104, tg = 52.65 t/s
  377. 3.23.819.357 I slot print_timing: id 0 | task 1824 | prompt eval time = 917.34 ms / 873 tokens ( 1.05 ms per token, 951.66 tokens per second)
  378. 3.23.819.363 I slot print_timing: id 0 | task 1824 | eval time = 2445.32 ms / 136 tokens ( 17.98 ms per token, 55.62 tokens per second)
  379. 3.23.819.365 I slot print_timing: id 0 | task 1824 | total time = 3362.66 ms / 1009 tokens
  380. 3.23.819.365 I slot print_timing: id 0 | task 1824 | graphs reused = 1642
  381. 3.23.819.367 I slot print_timing: id 0 | task 1824 | draft acceptance = 0.14091 ( 93 accepted / 660 generated)
  382. 3.23.819.368 I slot print_timing: id 0 | task 1824 | adaptive dm: fringe=0.00 n_max=15
  383. 3.23.819.397 I statistics dflash: #calls(b,g,a) = 16 1787 1787, #gen drafts = 1787, #acc drafts = 1257, #gen tokens = 26462, #acc tokens = 3896, dur(b,g,a) = 0.012, 13245.810, 0.563 ms
  384. 3.23.820.931 I slot release: id 0 | task 1824 | stop processing: n_tokens = 29587, truncated = 0
  385. 3.23.821.074 I srv update_slots: all slots are idle
  386. 3.24.189.897 I srv params_from_: Chat format: peg-native
  387. 3.24.238.227 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.990 (> 0.100 thold), f_keep = 0.996
  388. 3.24.238.232 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.990, keep=0.996, n_max=15)
  389. 3.24.238.697 I reasoning-budget: activated, budget=2147483647 tokens
  390. 3.24.238.995 I slot launch_slot_: id 0 | task 1873 | processing task, is_child = 0
  391. 3.24.239.096 I slot update_slots: id 0 | task 1873 | Checking checkpoint with [28576, 28576] against 29466...
  392. 3.24.314.858 W slot update_slots: id 0 | task 1873 | restored context checkpoint (pos_min = 28576, pos_max = 28576, n_tokens = 28577, n_past = 28577, size = 236.736 MiB)
  393. 3.24.319.144 W find_slot: non-consecutive token position 29088 after 28578 for sequence 0 with 512 new tokens
  394. 3.24.319.759 W find_slot: non-consecutive token position 29088 after 28578 for sequence 0 with 512 new tokens
  395. 3.25.498.203 I slot create_check: id 0 | task 1873 | created context checkpoint 9 of 32 (pos_min = 29751, pos_max = 29751, n_tokens = 29752, size = 249.626 MiB)
  396. 3.25.786.656 I reasoning-budget: deactivated (natural end)
  397. 3.27.247.207 I slot print_timing: id 0 | task 1873 | n_decoded = 100, tg = 59.09 t/s
  398. 3.30.249.838 I slot print_timing: id 0 | task 1873 | n_decoded = 246, tg = 52.40 t/s
  399. 3.30.301.705 I slot print_timing: id 0 | task 1873 | prompt eval time = 1314.91 ms / 1179 tokens ( 1.12 ms per token, 896.64 tokens per second)
  400. 3.30.301.711 I slot print_timing: id 0 | task 1873 | eval time = 4746.63 ms / 247 tokens ( 19.22 ms per token, 52.04 tokens per second)
  401. 3.30.301.713 I slot print_timing: id 0 | task 1873 | total time = 6061.54 ms / 1426 tokens
  402. 3.30.301.714 I slot print_timing: id 0 | task 1873 | graphs reused = 1733
  403. 3.30.301.715 I slot print_timing: id 0 | task 1873 | draft acceptance = 0.11111 ( 155 accepted / 1395 generated)
  404. 3.30.301.716 I slot print_timing: id 0 | task 1873 | adaptive dm: fringe=0.00 n_max=15
  405. 3.30.301.738 I statistics dflash: #calls(b,g,a) = 17 1880 1880, #gen drafts = 1880, #acc drafts = 1323, #gen tokens = 27857, #acc tokens = 4051, dur(b,g,a) = 0.013, 13951.177, 0.582 ms
  406. 3.30.303.309 I slot release: id 0 | task 1873 | stop processing: n_tokens = 30004, truncated = 0
  407. 3.30.303.431 I srv update_slots: all slots are idle
  408.  
  409. total : 2.29.6s (same speed??)
Add Comment
Please, Sign In to add comment
Public Pastes
We use cookies for various purposes including analytics. By continuing to use Pastebin, you agree to our use of cookies as described in the Cookies Policy.  OK, I Understand
Not a member of Pastebin yet?
Sign Up, it unlocks many cool features!