Using gpu device 0: GRID K520 (CNMeM is disabled)
using CUDNN backend
{'aug_params': {'allow_stretch': True,
                'do_flip': True,
                'rotation_range': (0, 360),
                'shear_range': (0, 0),
                'translation_range': (-40, 40),
                'zoom_range': (0.8695652173913044, 1.15)},
 'balance_ratio': 0.975,
 'balance_weights': array([  1.36094537,  14.3782235 ,   6.63756614,  40.23596793,  49.61299435]),
 'batch_size_test': 128,
 'batch_size_train': 128,
 'final_balance_weights': array([ 1.,  2.,  2.,  2.,  2.]),
 'h': 112,
 'name': 'c_128_5x5_32',
 'schedule': {0: 0.003, 150: 0.0003, 201: 'stop'},
 'sigma': 0.5,
 'test_dir': 'data/test_tiny',
 'train_dir': 'data/train_tiny',
 'w': 112,
 'weight_decay': 0.0005}
(35126, 35126, 35126)
/home/ubuntu/dataset/kaggle_diabetic/solution/local/lib/python2.7/site-packages/lasagne/init.py:86: UserWarning: The uniform initializer no longer uses Glorot et al.'s approach to determine the bounds, but defaults to the range (-0.01, 0.01) instead. Please use the new GlorotUniform initializer to get the old behavior. GlorotUniform is now the default for all layers.
  warnings.warn("The uniform initializer no longer uses Glorot et al.'s "
/home/ubuntu/dataset/kaggle_diabetic/nn.py:147: UserWarning: lasagne.objectives.Objective is deprecated and will be removed for the first release of Lasagne. For alternatives, please see: http://lasagne.readthedocs.org/en/latest/modules/objectives.html
  obj = objective(output_layer, **objective_params)
/home/ubuntu/dataset/kaggle_diabetic/solution/local/lib/python2.7/site-packages/lasagne/layers/helper.py:69: UserWarning: get_all_layers() has been changed to return layers in topological order. The former implementation is still available as get_all_layers_old(), but will be removed before the first release of Lasagne. To ignore this warning, use `warnings.filterwarnings('ignore', '.*topo.*')`.
  warnings.warn("get_all_layers() has been changed to return layers in "
/home/ubuntu/dataset/kaggle_diabetic/solution/local/lib/python2.7/site-packages/lasagne/objectives.py:241: UserWarning: lasagne.objectives.mse() is deprecated and will be removed for the first release of Lasagne. Use lasagne.objectives.squared_error() instead.
  losses = self.loss_function(network_output, target)
couldn't load weights starting from scratch
fitting ...
data/train_tiny/40267_left.tiff
# Neural Network with 1808481 learnable parameters

## Layer information

name           size         total    cap.Y    cap.X    cov.Y    cov.X
-------------  ---------  -------  -------  -------  -------  -------
input0         3x112x112    37632   100.00   100.00   100.00   100.00
conv2ddnn1     32x56x56    100352   100.00   100.00     4.46     4.46
conv2ddnn2     32x56x56    100352    42.86    42.86     6.25     6.25
maxpool2ddnn3  32x27x27     23328    42.86    42.86     6.25     6.25
conv2ddnn4     64x14x14     12544    78.95    78.95    16.96    16.96
conv2ddnn5     64x14x14     12544    36.00    36.00    22.32    22.32
conv2ddnn6     64x14x14     12544    29.03    29.03    27.68    27.68
maxpool2ddnn7  64x6x6        2304    29.03    29.03    27.68    27.68
conv2ddnn8     128x6x6       4608    55.10    55.10    43.75    43.75
conv2ddnn9     128x6x6       4608    40.30    40.30    59.82    59.82
conv2ddnn10    128x6x6       4608    31.76    31.76    75.89    75.89
rmspool11      128x2x2        512   100.00   100.00   100.00   100.00
dropout12      128x2x2        512   100.00   100.00   100.00   100.00
dense13        1024          1024   100.00   100.00   100.00   100.00
featurepool14  512            512   100.00   100.00   100.00   100.00
dropout15      512            512   100.00   100.00   100.00   100.00
dense16        1024          1024   100.00   100.00   100.00   100.00
featurepool17  512            512   100.00   100.00   100.00   100.00
dense18        1                1   100.00   100.00   100.00   100.00

Explanation
    X, Y:    image dimensions
    cap.:    learning capacity
    cov.:    coverage of image
    magenta: capacity too low (<1/6)
    cyan:    image coverage too high (>100%)
    red:     capacity too low and coverage too high


Traceback (most recent call last):
  File "train_nn.py", line 41, in <module>
    main()
  File "/home/ubuntu/dataset/kaggle_diabetic/solution/local/lib/python2.7/site-packages/click/core.py", line 716, in __call__
    return self.main(*args, **kwargs)
  File "/home/ubuntu/dataset/kaggle_diabetic/solution/local/lib/python2.7/site-packages/click/core.py", line 696, in main
    rv = self.invoke(ctx)
  File "/home/ubuntu/dataset/kaggle_diabetic/solution/local/lib/python2.7/site-packages/click/core.py", line 889, in invoke
    return ctx.invoke(self.callback, **ctx.params)
  File "/home/ubuntu/dataset/kaggle_diabetic/solution/local/lib/python2.7/site-packages/click/core.py", line 534, in invoke
    return callback(*args, **kwargs)
  File "train_nn.py", line 38, in main
    net.fit(files, labels)
  File "/home/ubuntu/dataset/kaggle_diabetic/solution/src/nolearn-master/nolearn/lasagne/base.py", line 334, in fit
    self.train_loop(X, y)
  File "/home/ubuntu/dataset/kaggle_diabetic/nn.py", line 250, in train_loop
    raise ValueError("non finite loss")
ValueError: non finite loss