R
RunPod3w ago
yoyo

5090 error serverless

Does the vllm image have the old pytorch? im getting this NVIDIA GeForce RTX 5090 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_70 sm_75 sm_80 sm_86 sm_90. If you want to use the NVIDIA GeForce RTX 5090 GPU with PyTorch, please check the instructions at https://pytorch.org/get-started/locally/
PyTorch
Get Started
Set up PyTorch easily with local installation or supported cloud platforms.
6 Replies
Jason
Jason3w ago
You should use newer pytorch yeah, it hasn't been updated I guess
yoyo
yoyoOP3w ago
oh ok hmmm should i just change the image? of the vllm provided by runpod or what
Jason
Jason3w ago
I'm not sure if there is any Let me help check i think there is no image for cuda 12.8, you need to build it on your own
yoyo
yoyoOP3w ago
fine w me can you link it? oh wait there's no aight aight
Jason
Jason3w ago
i hope they build it soon
Dj
Dj3w ago
Working on this :) I needed to spend some time on our containers repo this worker will start depending on

Did you find this page helpful?