RuntimeError: cuDNN error: CUDNN_STATUS_NOT_INITIALIZED using pytorch

Question

I am trying to run a simple pytorch sample code. It's works fine using CPU. But when using GPU, i get this error message:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py", line 889, in _call_impl
    result = self.forward(*input, **kwargs)
  File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/conv.py", line 263, in forward
    return self._conv_forward(input, self.weight, self.bias)
  File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/conv.py", line 260, in _conv_forward
    self.padding, self.dilation, self.groups)
RuntimeError: cuDNN error: CUDNN_STATUS_NOT_INITIALIZED

The code i am trying to run is the following:

import torch
from torch import nn
m = nn.Conv1d(16, 33, 3, stride=2)
m=m.to('cuda')
input = torch.randn(20, 16, 50)
input=input.to('cuda')
output = m(input)

I am running this code in a NVIDIA docker with CUDA version 10.2 and my GPU is a RTX 2070

One hint which is not related to your problem. Please do not use python keywords as a variable because this can cause some very ugly and difficult problems. — mosc9575, Mar 11 '21 at 20:05
I have exactly the same problem on CUDA 10.2. Did you solve it? — Guojun Zhang, Mar 15 '21 at 21:36
@GuojunZhang I solved it by using the pytorch container for nvidia docker. — Eduardo H, Mar 23 '21 at 19:59

creiser · Answer 1 · 2021-11-02T10:06:17.657

In my case it actually had nothing do with the PyTorch/CUDA/cuDNN version. PyTorch initializes cuDNN lazily whenever a convolution is executed for the first time. However, in my case there was not enough GPU memory left to initialize cuDNN because PyTorch itself already held the entire memory in its internal cache. One can release the cache manually with "torch.cuda.empty_cache()" right before the first convolution that is executed. A cleaner solution is to force cuDNN initialization at the beginning by doing a mock convolution:

def force_cudnn_initialization():
    s = 32
    dev = torch.device('cuda')
    torch.nn.functional.conv2d(torch.zeros(s, s, s, s, device=dev), torch.zeros(s, s, s, s, device=dev))

Calling the above function at the very beginning of the program solved the problem for me.

saturn660 · Answer 2 · 2021-03-21T03:51:21.340

21

There is some discussion regarding this here. I had the same issue but using cuda 11.1 resolved it for me.

This is the exact pip command

pip install torch==1.8.0+cu111 torchvision==0.9.0+cu111 torchaudio==0.8.0 -f https://download.pytorch.org/whl/torch_stable.html

edited Mar 21 '21 at 03:51

answered Mar 21 '21 at 03:44

saturn660

313
1
5

score 7 · Answer 3 · answered Mar 16 '21 at 09:09

I am also using Cuda 10.2. I had the exact same error when upgrading torch and torchvision to the latest version (torch-1.8.0 and torchvision-0.9.0). Which version are you using?

I guess this is not the best solution but by downgrading to torch-1.7.1 and torchvision-0.8.2 it works just fine.

score 2 · Answer 4 · answered Oct 26 '22 at 04:10

2

I had the same issue when I was training yolov7 with a chess dataset. By reducing batch size from 8 to 4, the issue was solved.

answered Oct 26 '22 at 04:10

Flying Fox

21
1

Your answer could be improved with additional supporting information. Please [edit] to add further details, such as citations or documentation, so that others can confirm that your answer is correct. You can find more information on how to write good answers [in the help center](/help/how-to-answer). – Community Oct 29 '22 at 18:22
As noted, some additional information would be helpful. – craigb Oct 30 '22 at 20:53

score 1 · Answer 5 · answered Dec 03 '21 at 08:34

In my cases this error occurred when trying to estimate loss. I used a mixed bce-dice loss. It turned out that my output was linear instead of sigmoid. I then used the sigmoid predictions as of bellow and worked fine.

output = torch.nn.Sigmoid()(output)
loss = criterion1(output, target)

score 1 · Answer 6 · answered Dec 02 '22 at 17:36

In my case, I had an array indexing operation but the index was out of bounds. CUDA did not tell me that. I was using inference on a neural network. So I moved to CPU instead of the GPU. The logs were much more informative after that. For debugging if you see this error, switch to CPU first and you will know what to do.

score 1 · Answer 7 · answered Dec 09 '22 at 09:23

1

In my problem i used to kill exisiting process in gpu.Use nvidia-smi to check what are the process are running.Use killall -9 python3(what process you want) to kill process.After freeup space then run the process.

answered Dec 09 '22 at 09:23

sathish kumar

11
2

score 0 · Answer 8 · answered Jul 21 '23 at 08:33

0

Sometimes, if any error happens in the CUDA c++ code that is converted into .so file and used inside Python code, it could cause this problem, so check your C++ source code if you have any.

answered Jul 21 '23 at 08:33

Tony.sy

11
2

RuntimeError: cuDNN error: CUDNN_STATUS_NOT_INITIALIZED using pytorch

8 Answers8

Linked