Closed sky-fly97 closed 1 year ago
Please check your CUDA Version, Graphics card type, and corresponding torch version. We can successfully compile on V100, 3090, and A100 with CUDA >= 11.0.
I have the same question,my version is torch==1.10.0+cu111,then came the same mistake. I also try the torch==1.9.1+cu111,then I was able to run python setup.py develop successfully.My gpu is 3090, cuda version 11.1
I have the same question,my version is torch==1.10.0+cu111,then came the same mistake. I also try the torch==1.9.1+cu111,then I was able to run python setup.py develop successfully.My gpu is 3090, cuda version 11.1
Yes, I also try the torch==1.9.1+cu111 and it works, but can you works when training?
File "/nvme/anaconda3/envs/test/lib/python3.9/site-packages/torch/nn/modules/conv.py", line 439, in _conv_forward
return F.conv2d(input, weight, bias, self.stride,
RuntimeError: cuDNN error: CUDNN_STATUS_NOT_INITIALIZED
One of the environments we succeed in is torch 1.9.1+cu111, CUDA >11.1, V100.
I have the same question,my version is torch==1.10.0+cu111,then came the same mistake. I also try the torch==1.9.1+cu111,then I was able to run python setup.py develop successfully.My gpu is 3090, cuda version 11.1
Yes, I also try the torch==1.9.1+cu111 and it works, but can you works when training?
File "/nvme/anaconda3/envs/test/lib/python3.9/site-packages/torch/nn/modules/conv.py", line 439, in _conv_forward return F.conv2d(input, weight, bias, self.stride, RuntimeError: cuDNN error: CUDNN_STATUS_NOT_INITIALIZED
This problem does not seem to be our problem. It may be that you have problems installing torch 1.9.1 in your environment.
We build a new environment with torch==1.10.0+cu111 and torchvision==0.11.0+cu111 and compile successfully.
I have the same question,my version is torch==1.10.0+cu111,then came the same mistake. I also try the torch==1.9.1+cu111,then I was able to run python setup.py develop successfully.My gpu is 3090, cuda version 11.1
Yes, I also try the torch==1.9.1+cu111 and it works, but can you works when training?
File "/nvme/anaconda3/envs/test/lib/python3.9/site-packages/torch/nn/modules/conv.py", line 439, in _conv_forward return F.conv2d(input, weight, bias, self.stride, RuntimeError: cuDNN error: CUDNN_STATUS_NOT_INITIALIZED
Has the problem been solved?
I have the same question,my version is torch==1.10.0+cu111,then came the same mistake. I also try the torch==1.9.1+cu111,then I was able to run python setup.py develop successfully.My gpu is 3090, cuda version 11.1
Yes, I also try the torch==1.9.1+cu111 and it works, but can you works when training?
File "/nvme/anaconda3/envs/test/lib/python3.9/site-packages/torch/nn/modules/conv.py", line 439, in _conv_forward return F.conv2d(input, weight, bias, self.stride, RuntimeError: cuDNN error: CUDNN_STATUS_NOT_INITIALIZED
Has the problem been solved?
Yes, I can train it successfully on torch==1.9.1+cu111!
I have the same question,my version is torch==1.10.0+cu111,then came the same mistake. I also try the torch==1.9.1+cu111,then I was able to run python setup.py develop successfully.My gpu is 3090, cuda version 11.1
Yes, I also try the torch==1.9.1+cu111 and it works, but can you works when training?
File "/nvme/anaconda3/envs/test/lib/python3.9/site-packages/torch/nn/modules/conv.py", line 439, in _conv_forward return F.conv2d(input, weight, bias, self.stride, RuntimeError: cuDNN error: CUDNN_STATUS_NOT_INITIALIZED
Has the problem been solved?
Yes, I can train it successfully on torch==1.9.1+cu111!
Great! Hope everything goes well!
Hi, my pytorch version used to be 1.8.1, then I was able to run
python setup.py develop
successfully. But the required version is greater than 1.9. so I created a new environment,pip install torch==1.10.0+cu111 torchvision==0.11.0+cu111 torchaudio==0.10.0 -f https://download.pytorch.org/whl/torch_stable.html
, and had problems rompiling, it looks like it's inpcdet/ops/ingroup_inds