CUDNN_STATUS_NOT_SUPPORTED error

1

I am getting this error I have read forum posts tackling this error message but mine seems to be a different issue.
Running pytorchv1 on cuda 10.0 with cudnn 7.4.1

Traceback (most recent call last):
File “depth_train.py”, line 415, in
main()
File “depth_train.py”, line 373, in main
loss = train(imgL_crop, imgR_crop, disp_crop_L)
File “depth_train.py”, line 298, in train
loss.backward()
File “/home/pranay/anaconda3/lib/python3.6/site-packages/torch/tensor.py”, line 106, in backward
torch.autograd.backward(self, gradient, retain_graph, create_graph)
File “/home/pranay/anaconda3/lib/python3.6/site-packages/torch/autograd/init.py”, line 93, in backward
allow_unreachable=True) # allow_unreachable flag
RuntimeError: cuDNN error: CUDNN_STATUS_NOT_SUPPORTED. This error may appear if you passed in a non-contiguous input.

2

Hi, can you share your training script in question?

3

I am afraid I can’t. It would be too long anyway.