My configuration is Keras + tensorflow backend. I'm getting the following errors:
```
STDERR: E tensorflow/stream_executor/cuda/cuda_dnn.cc:385] could not create cudnn handle: CUDNN_STATUS_INTERNAL_ERROR
STDERR: E tensorflow/stream_executor/cuda/cuda_dnn.cc:352] could not destroy cudnn handle: CUDNN_STATUS_BAD_PARAM
STDERR: F tensorflow/core/kernels/conv_ops.cc:532] Check failed: stream->parent()->GetConvolveAlgorithms(&algorithms)
STDOUT: Epoch 1/10
STDERR: /train.sh: line 51: 7 Aborted (core dumped)
```
This is very strange. It doesn't seem to have anything to do with my code. Please suggest.
Created by Li Shen thefaculty It's OK now. I reduced the batch size. It seems to be related with insufficient GPU memory but it didn't show so. Hi Li,
Is this an error that you can reproduce on the Express Lane? If yes, can you please provide us with the submission IDs as well as the version of CUDA and cuDNN that you are using?
Thanks!