Hi everyone,
I am trying to implement waveUnet in fastai. A very good implemention in tensorflow is already done at https://github.com/f90/Wave-U-Net . I am trying to reproduce the results mentioned in paper; but when I run the repo’s scripts I get the following errors:
2019-07-25 05:51:38.470459: I tensorflow/core/kernels/data/shuffle_dataset_op.cc:94] Filling up shuffle buffer (this may take a while): 2001 of 8000
2019-07-25 05:51:48.423484: I tensorflow/core/kernels/data/shuffle_dataset_op.cc:94] Filling up shuffle buffer (this may take a while): 4201 of 8000
2019-07-25 05:51:58.262603: I tensorflow/core/kernels/data/shuffle_dataset_op.cc:94] Filling up shuffle buffer (this may take a while): 6448 of 8000
2019-07-25 05:52:04.644012: I tensorflow/core/kernels/data/shuffle_dataset_op.cc:129] Shuffle buffer filled.
2019-07-25 05:52:04.890160: E tensorflow/stream_executor/cuda/cuda_dnn.cc:455] could not create cudnn handle: CUDNN_STATUS_NOT_INITIALIZED
2019-07-25 05:52:04.890268: E tensorflow/stream_executor/cuda/cuda_dnn.cc:463] possibly insufficient driver version: 384.183.0
2019-07-25 05:52:04.890282: E tensorflow/stream_executor/cuda/cuda_dnn.cc:427] could not destroy cudnn handle: CUDNN_STATUS_BAD_PARAM
2019-07-25 05:52:04.890291: F tensorflow/core/kernels/conv_ops.cc:713] Check failed: stream->parent()->GetConvolveAlgorithms( conv_parameters.ShouldIncludeWino
gradNonfusedAlgo<T>(), &algorithms)
Aborted (core dumped)
I am using tesnorflow-gpu==1.8.0, cudnn=7.0.5, cuda=9.1.84, ubuntu=16.04.
Please help.
Thank you