RuntimeError: CUDNN_STATUS_INTERNAL_ERROR

1

Traceback (most recent call last):
File “main-gpu.py”, line 72, in
outputs = cnn(images)
File “/home/luke/virtualpytorch/venv/local/lib/python2.7/site-packages/torch/nn/modules/module.py”, line 224, in call
result = self.forward(*input, **kwargs)
File “main-gpu.py”, line 49, in forward
out = self.layer1(x)
File “/home/luke/virtualpytorch/venv/local/lib/python2.7/site-packages/torch/nn/modules/module.py”, line 224, in call
result = self.forward(*input, **kwargs)
File “/home/luke/virtualpytorch/venv/local/lib/python2.7/site-packages/torch/nn/modules/container.py”, line 67, in forward
input = module(input)
File “/home/luke/virtualpytorch/venv/local/lib/python2.7/site-packages/torch/nn/modules/module.py”, line 224, in call
result = self.forward(*input, **kwargs)
File “/home/luke/virtualpytorch/venv/local/lib/python2.7/site-packages/torch/nn/modules/conv.py”, line 254, in forward
self.padding, self.dilation, self.groups)
File “/home/luke/virtualpytorch/venv/local/lib/python2.7/site-packages/torch/nn/functional.py”, line 52, in conv2d
return f(input, weight, bias)

I trapped in this problem for 3 days,is someone know how to solve it?

2

Could you provide the input weight and bias shapes of the specific conv layer please?

Also, we improved shape checking in conv since 0.2. So if you update to master / wait for the coming 0.3, it might be clearer.

3

I got my problem solved by applying the solution: https://stackoverflow.com/questions/45810356/runtimeerror-cudnn-status-internal-error

4

Thank you very much.I solved the problem in your way.