Bug Report No Version Of Pytorch for cuda 12.3

[tomas122727]

If you have a question or would like help and support, please ask at our
forums.

If you are submitting a feature request, please preface the title with [feature request].
If you are submitting a bug report, please fill in the following details.

Issue description

since updating both my gpu driver and cuda to cuda 12.3 pytorch isnt detecting my gpu and is complaing that cuda is not availible and that i dont have supported gpu when it was working just fine my gpu is a rtx 3060 yes with cuda support

Code example

torch.is.availible()

System Info

Please copy and paste the output from our
environment collection script
(or fill out the checklist below manually).

You can get the script and run it with:

wget https://raw.githubusercontent.com/pytorch/pytorch/main/torch/utils/collect_env.py
# For security purposes, please check the contents of collect_env.py before running it.
python collect_env.py

• PyTorch or Caffe2: pytorch
• How you installed PyTorch (conda, pip, source):
• Build command you used (if compiling from source):
• OS: windows 10
• PyTorch version: latest available
• Python version:
• CUDA/cuDNN version: latest
• GPU models and configuration:
• GCC version (if compiling from source):
• CMake version:
• Versions of any other relevant libraries: cuda toolkit 12.3

[tomas122727]

torch returns that no cuda gpu is availible when it was working fine before ?

[malfet]

Do you mind opening another issue and fill in necessary details, namely:

• Your OS
• How do you install PyTorch (pip/conda/build from source)
• Output of python3 -mtorch.utils.collect_env

But from the look of it, this sounds like a local system setup problem, perhaps https://discuss.pytorch.org/ is a more appropriate place to ask such questions?

[sascharo]

Does PyTorch work with CUDA 12.3? Isn't 12.1 the last version supported?

[tomas122727]

py torch only has 12.1 support ?

[tomas122727]

Versions of relevant libraries:
[pip3] numpy==1.24.1
[pip3] torch==2.1.0+cu118
[pip3] torchaudio==2.1.0+cu118
[pip3] torchvision==0.16.0+cu118
[conda] blas 1.0 mkl
[conda] mkl 2023.1.0 h6b88ed4_46358
[conda] mkl-service 2.4.0 py311h2bbff1b_1
[conda] mkl_fft 1.3.8 py311h2bbff1b_0
[conda] mkl_random 1.2.4 py311h59b6b97_0
[conda] numpy 1.26.1 pypi_0 pypi
[conda] numpy-base 1.26.0 py311hd01c5d8_0
[conda] pytorch 2.1.1 py3.11_cuda11.8_cudnn8_0 pytorch
[conda] pytorch-cuda 11.8 h24eeafa_5 pytorch
[conda] pytorch-mutex 1.0 cuda pytorch
[conda] torch 2.0.0+cu118 pypi_0 pypi
[conda] torchaudio 2.1.1 pypi_0 pypi
[conda] torchvision 0.16.1 pypi_0 pypi

[tomas122727]

note i downgraded to cuda 11.8

[tomas122727]

but still things arent recognized no matter how i installed pytorch via conda or pip it still compains

[sascharo]

py torch only has 12.1 support ?

https://discuss.pytorch.org/t/question-i-have-a-question-about-installing-pytorch/191829/2

"PyTorch is compatible with CUDA 12.3 and will use your locally installed CUDA toolkit for source builds. The binaries ship with their own CUDA dependencies, won’t use your local CUDA toolkit, and only a properly installed NVIDIA driver is needed."

[jobyclash2]

i have same issue

[boyXiong]

i have same issue, don't use gpu to work

[AmoghTantradi]

same issue!

[thinkhp]

Instead of using conda install ..., try using pip install torch torchvision torchaudio. It works for me.
My environment: WSL with Ubuntu, CUDA 12.3, Python 3.10.

[sascharo]

Instead of using conda install ..., try using pip install torch torchvision torchaudio. It works for me.
My environment: WSL with Ubuntu, CUDA 12.3, Python 3.10.

That's good for you. It didn't work for me with WSL2 Ubuntu 22.04.2 and Python 3.10. I tried both, conda and pip3. I could only get it to work with CUDA 12.1 (driver version 531.14). With driver versions that came with CUDA 12.3, PyTorch didn't recognize the CUDA device.