Some layers do not support non-cuDNN kernels because they do not meet the criteria.

[vovasilenko]

You choose NUTLS-LSTM... WARNING:tensorflow:Layer msfe6_en_lstm will not use cuDNN kernels since it doesn't meet the criteria. It will use a generic GPU kernel as fallback when running on GPU. WARNING:tensorflow:Layer msfe5_en_lstm will not use cuDNN kernels since it doesn't meet the criteria. It will use a generic GPU kernel as fallback when running on GPU. WARNING:tensorflow:Layer msfe4_en_lstm will not use cuDNN kernels since it doesn't meet the criteria. It will use a generic GPU kernel as fallback when running on GPU. WARNING:tensorflow:Layer msfe4_en2_lstm will not use cuDNN kernels since it doesn't meet the criteria. It will use a generic GPU kernel as fallback when running on GPU. WARNING:tensorflow:Layer msfe4_en3_lstm will not use cuDNN kernels since it doesn't meet the criteria. It will use a generic GPU kernel as fallback when running on GPU. WARNING:tensorflow:Layer msfe3_en_lstm will not use cuDNN kernels since it doesn't meet the criteria. It will use a generic GPU kernel as fallback when running on GPU. WARNING:tensorflow:Layer lstm will not use cuDNN kernels since it doesn't meet the criteria. It will use a generic GPU kernel as fallback when running on GPU. WARNING:tensorflow:Layer msfe3_de_lstm will not use cuDNN kernels since it doesn't meet the criteria. It will use a generic GPU kernel as fallback when running on GPU. WARNING:tensorflow:Layer msfe4_de_lstm will not use cuDNN kernels since it doesn't meet the criteria. It will use a generic GPU kernel as fallback when running on GPU. WARNING:tensorflow:Layer msfe4_de2_lstm will not use cuDNN kernels since it doesn't meet the criteria. It will use a generic GPU kernel as fallback when running on GPU. WARNING:tensorflow:Layer msfe4_de3_lstm will not use cuDNN kernels since it doesn't meet the criteria. It will use a generic GPU kernel as fallback when running on GPU. WARNING:tensorflow:Layer msfe5_de_lstm will not use cuDNN kernels since it doesn't meet the criteria. It will use a generic GPU kernel as fallback when running on GPU. WARNING:tensorflow:Layer msfe6_de_lstm will not use cuDNN kernels since it doesn't meet the criteria. It will use a generic GPU kernel as fallback when running on GPU.

Were there any similar messages during training?

How can this problem be solved?

Perhaps because of these problems, I get the following error: [tensorflow/stream_executor/cuda/cuda_fft.cc:439] failed to initialize batched cufft plan with customized allocator: Failed to make cuFFT batched plan. Aborted (core dumped).