deep-learning-models icon indicating copy to clipboard operation
deep-learning-models copied to clipboard

Published 20 hours ago • verified publisher icon fchollet

Resnet50 fails on TensorFlow-gpu

Open axelrodR opened this issue 8 years ago • 0 comments

The Resnet50 example fails on TensorFlow-GPU. The example works well on the non GPU TensorFlow.

It fails on load_weights. Parts of the traceback:

File "C:/Users/gslab/Downloads/resnet50.py", line 251, in ResNet50 model.load_weights(weights_path)

File "C:\Program Files\Anaconda3\lib\site-packages\keras\engine\topology.py", line 2495, in load_weights load_weights_from_hdf5_group(f, self.layers)

File "C:\Program Files\Anaconda3\lib\site-packages\keras\engine\topology.py", line 2908, in load_weights_from_hdf5_group K.batch_set_value(weight_value_tuples)

<<<< --- long list of calls in tensorflow_backend.py --->>

File "C:\Program Files\Anaconda3\lib\site-packages\tensorflow\python\client\session.py", line 420, in init self._assert_fetchable(graph, fetch)

File "C:\Program Files\Anaconda3\lib\site-packages\tensorflow\python\client\session.py", line 437, in _assert_fetchable 'Operation %r has been marked as not fetchable.' % op.name)

ValueError: Operation 'init_2' has been marked as not fetchable.

axelrodR avatar Jul 12 '17 17:07 axelrodR