You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I have built the docker image from the provided Dockerfile with cuda11.6. Now I am trying to run models on an H100 GPU. But it seems the corresponding PyTorch version is not compatible with the CUDA capability. As shown below:
/usr/local/lib/python3.8/dist-packages/torch/cuda/__init__.py:155: UserWarning:
NVIDIA H100 80GB HBM3 with CUDA capability sm_90 is not compatible with the current PyTorch installation.
The current PyTorch install supports CUDA capabilities sm_37 sm_50 sm_60 sm_70 sm_75 sm_80 sm_86.
If you want to use the NVIDIA H100 80GB HBM3 GPU with PyTorch, please check the instructions at https://pytorch.org/get-started/locally/
warnings.warn(incompatible_device_warn.format(device_name, capability, " ".join(arch_list), device_name))
Do we have a newer version of Dockerfile which is compatible? Thanks a lot in advance for the support!
The text was updated successfully, but these errors were encountered:
Hi!
I have built the docker image from the provided Dockerfile with cuda11.6. Now I am trying to run models on an H100 GPU. But it seems the corresponding PyTorch version is not compatible with the CUDA capability. As shown below:
Do we have a newer version of Dockerfile which is compatible? Thanks a lot in advance for the support!
The text was updated successfully, but these errors were encountered: