Problem with cuda/pytorch on Fedora

1

I have a Fedora 41 Linux and i want to use pytorch with cuda. Everywhere I look i see that info that pytorch works witch cuda up to 12.5 but i have cuda 12.7. What can I do? That’s what nvidia-smi returns for me:

Wed Dec 11 20:04:25 2024
+
| NVIDIA-SMI 565.57.01 Driver Version: 565.57.01 CUDA Version: 12.7 |
|
| GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |

| 0 NVIDIA GeForce GTX 1650 Off | 00000000:01:00.0 Off | N/A |
| N/A 36C P8 3W / 50W | 1MiB / 4096MiB | 0% Default |
| | | N/A |
+

It would be best if i could get it for python 3.13 if that’s possible.

I’m quite new to linux and cuda so I would apreciate some help

2

PyTorch works with all released CUDA toolkits up to 12.6U3, which is the latest CUDA toolkit release, and the PyTorch nightly binaries already ship with it.
The nvidia-smi output shows the driver, which is also supported as it’s new enough.

3

To my understanding it the output also contains CUDA version: 2.7. Is there any way to downgrade the version to 2.6 so that I can use pytorch?