RuntimeError: cuDNN error: CUDNN_STATUS_INTERNAL_ERROR

To solve PyTorch Model Training: RuntimeError: cuDNN error: CUDNN_STATUS_INTERNAL_ERROR follow any of the below methods.

Error log

 self.dropout, self.training, self.bidirectional, self.batch_first)
else:

RuntimeError: cuDNN error: CUDNN_STATUS_INTERNAL_ERROR
or
RuntimeError: CUDA error: unspecified launch failure

How to solve RuntimeError: cuDNN error: CUDNN_STATUS_INTERNAL_ERROR ?

The error RuntimeError: cuDNN error: CUDNN STATUS INTERNAL ERROR is generally tough to troubleshoot, however it’s generally due to a lack of memory. Normally, you would receive an out of memory error, however depending on where it occurs, PyTorch is unable to intercept the error and hence cannot deliver a relevant error message.

RuntimeError: cuDNN error: CUDNN_STATUS_INTERNAL_ERROR

Try reducing the number of workers and also try to downgrade the cudatoolkit to version 10.1. So try to reinstall pytorch with cudatoolkit 10.1 using :

conda install pytorch torchvision cudatoolkit=10.1

Hope it helps.

Also Read : Node/Postgresql Error: connect ECONNREFUSED 127.0.0.1:5432 at TCPConnectWrap.afterConnect [as oncomplete] (net.js:1141:16)