cuDNN error

[liuyanice]

which version of pytorch is used, could you share the package lists?

[Qsingle]

Good issue, we use torch 1.13.1+CU117. It seems that the model or data can not be feed into the GPU. I suggest you check the version of CUDA.

[liuyanice]

my issue： torch 1.13.1+CU117 nvidia-smi: CUDA Version:11.7 Driver Version:515.105.01 nvcc: cuda_11.3.r11.3

cuDNN error: CUDNN_STATUS_NOT_INITIALIZED

[Qsingle]

Can you tell me the GPU you use? The version of cudnn. If you use Anaconda or Minianaconda I suggest you to create one virtual environment by the following command. Or you can install the CUDA 11.7 and Cudnn 8.5.0.

conda create -n pytorch python=3.8 pytorch==1.13.1 torchvision==0.14.1 torchaudio==0.13.1 pytorch-cuda=11.7 -c pytorch -c nvidia
conda activate pytorch #activate the enviroment

[liuyanice]

RTX TI3090 single GPU, can you tell me the GPU you use?

[liuyanice]

[liuyanice]

[liuyanice]

[Qsingle]

I know the reason for this error. Similar to #6, your label may contain an invalid value like 255.

[liuyanice]

Can you share your datasets, I want to reproduce the results.

[liuyanice]

Beacause I have set num classes is 2 and used the blood vessel datasets. It also displays "cudnn error".

[liuyanice]

Thank you, I have already solved this problem.

[liuyanice]

It is indeed caused by an invalid value