RuntimeError: Expected tensor for argument #1 'input' to have the same device as tensor for argument #2 'weight'; but device 1 does not equal 0 (while checking arguments for cudnn_convolution)

Hello,
When I used multiple gpus to train model, I usually get this error. `RuntimeError: Expected tensor for argument #1 'input' to have the same device as tensor for argument #2 'weight'; but device 1 does not equal 0 (while checking arguments for cudnn_convolution)`
Could you tell me how to sole it.

best,
Honggu Liu

Reproduce the reported multi-GPU training failure and inspect device placement at the cudnn_convolution call. Done requires identifying the input/weight device mismatch and recording a reproducible resolution; the issue names no source file or test.

Written by the indexing model from the issue text.