My RTX5080 GPU can't work with PyTorch

1

\Local\Programs\Python\Python312\Lib\site-packages\torch\cuda_init_.py:235: UserWarning:
NVIDIA GeForce RTX 5080 with CUDA capability sm_120 is not compatible with the current PyTorch installation.
The current PyTorch install supports CUDA capabilities sm_50 sm_60 sm_61 sm_70 sm_75 sm_80 sm_86 sm_90.
If you want to use the NVIDIA GeForce RTX 5080 GPU with PyTorch, please check the instructions at Start Locally | PyTorch

and i tried Start Locally | PyTorch and find nothing to help
if there is an update suite my GPU
please email me

2

You need to install the nightly binaries with CUDA 12.8 (select nightly in the install matrix and replace the CUDA version in the URL with cu128). We’ll update the install matrix once all builds are ready.

3

When will the new version of pytorch be released?

4

Nightly binaries are released every night. 2.7 is planned for April I believe.

5

pip3 install --pre torchvision torchaudio --index-url /whl/nightly/cu128
Looking in indexes: /whl/nightly/cu128
Collecting torchvision
Using cached https://download.pytorch.org/whl/nightly/cu128/torchvision-0.22.0.dev20250302%2Bcu128-cp312-cp312-win_amd64.whl.metadata (6.3 kB)
ERROR: Could not find a version that satisfies the requirement torchaudio (from versions: none)

2 Likes
6

hi rtx5090 Error creating song: CUDA error: no kernel image is available for execution on the device Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.

7

use preview version
pip3 install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/cu128
if you can‘t match the day package，check version in https://download.pytorch.org/whl/torch/

8

how i can use preview version
pip3 install --pre torch torchvision torchaudio --index-url

9

“pip3 install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/cu128”
put url end of pip3 cmd

10

i can use anydesk fix my bc
thank you

11

hi
ERROR: Cannot install torch, torchaudio==2.6.0 and torchvision==0.22.0.dev20250226+cu128 because these package versions have conflicting dependencies.

12

https://download.pytorch.org/whl/nightly/torchvision/
find your version in this url
i use torchvision-0.22.0.dev20250302+cu128-cp312-cp312-win_amd64.whl
choose your version with +cu128, cpu version，os type,cpu type

13

i can’t find torchaudio-2.6.0.dev20250302+cu128-cp312-cp312-win_amd64.whl
i need this version

14
16

i got torchvision-0.22.0.dev20250306+cu128-cp312-cp312-win_amd64.whl
but Requires-Dist: torch (==2.7.0.dev20250305+cu128)
i edit 0305 to 0306 and install completed
if there is a bug 0306 requires 0305,you should know

17

torchaudio binaries for Windows are unblocked now with a workaround.

18

hi rtx 5090
pip3 install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/cu128 If you download the program
replay
Nvidia rvc
it will work because I tried but it didn’t work