DengPingFan/PraNet

RuntimeError: cuDNN error: CUDNN_STATUS_EXECUTION_FAILED

Closed this issue · 4 comments

evil-lov commented

Traceback (most recent call last):
File "/home/PraNet-master/MyTrain.py", line 116, in
train(train_loader, model, optimizer, epoch)
File "/home/PraNet-master/MyTrain.py", line 50, in train
loss.backward()
File "/home/.virtualenvs/pytor1.1/lib/python3.6/site-packages/torch/tensor.py", line 107, in backward
torch.autograd.backward(self, gradient, retain_graph, create_graph)
File "/home/.virtualenvs/pytor1.1/lib/python3.6/site-packages/torch/autograd/init.py", line 93, in backward
allow_unreachable=True) # allow_unreachable flag
RuntimeError: cuDNN error: CUDNN_STATUS_EXECUTION_FAILED

Process finished with exit code 1

evil-lov commented

Traceback (most recent call last):
File "/home/PraNet-master/MyTrain.py", line 116, in
train(train_loader, model, optimizer, epoch)
File "/home/PraNet-master/MyTrain.py", line 50, in train
loss.backward()
File "/home/.virtualenvs/pytor1.1/lib/python3.6/site-packages/torch/tensor.py", line 107, in backward
torch.autograd.backward(self, gradient, retain_graph, create_graph)
File "/home/.virtualenvs/pytor1.1/lib/python3.6/site-packages/torch/autograd/init.py", line 93, in backward
allow_unreachable=True) # allow_unreachable flag
RuntimeError: cuDNN error: CUDNN_STATUS_EXECUTION_FAILED

Process finished with exit code 1

cuda10.0+cudnn7.6.5+pytorch1.1

zihuanqiu commented

reboot your computer may help?

evil-lov commented

reboot your computer may help?
No, I finally chose to turn cudnn off, but it will be much slower

GewelsJI commented

I suggest you can install it via conda toolbox, which can introduce the correct dependencies of GPU automatically.