CRNN_Tensorflow: NotFoundError

I try to run the demo for one image, but a notfounderror ocurr . What should I do?

NotFoundError (see above for traceback): Restoring from checkpoint failed. This is most likely due to a Variable name or other graph key that is missing from the checkpoint. Please ensure that you have not altered the graph expected based on the checkpoint. Original error:

Key shadow/batch_normalization/beta not found in checkpoint
	 [[Node: save/RestoreV2 = RestoreV2[dtypes=[DT_FLOAT, DT_FLOAT, DT_FLOAT, DT_FLOAT, DT_FLOAT, ..., DT_FLOAT, DT_FLOAT, DT_FLOAT, DT_FLOAT, DT_FLOAT], _device="/job:localhost/replica:0/task:0/device:CPU:0"](_arg_save/Const_0_0, save/RestoreV2/tensor_names, save/RestoreV2/shape_and_slices)]]
	 [[Node: save/RestoreV2/_3 = _Recv[client_terminated=false, recv_device="/job:localhost/replica:0/task:0/device:GPU:0", send_device="/job:localhost/replica:0/task:0/device:CPU:0", send_device_incarnation=1, tensor_name="edge_7_save/RestoreV2", tensor_type=DT_FLOAT, _device="/job:localhost/replica:0/task:0/device:GPU:0"]()]]

About this issue

  • Original URL
  • State: closed
  • Created 6 years ago
  • Comments: 22 (17 by maintainers)

Most upvoted comments

I think this is due to the recent upgrade to TF 1.10. The batch normalization layers are now using tf.layers.batch_normalization, instead of tf.contrib.layers.batch_norm. This means that the checkpoint in the repo is no longer valid for the current code! I hadn’t thought of this as I submitted the PR. Maybe we should undo that change (and any having the same consequence)?

(And… that’s why one always wants to have regression tests! 😉