Issue while PyTorch installation on cuda 12.0

Hello PyTorch community,

I’m encountering an issue while trying to run PyTorch on my GPU, and I’m seeking some guidance on how to resolve it.
Problem Description:

I’m attempting to run a PyTorch script on my GPU, but I’m consistently getting the following error and warning:
Warning : H100 80GB HBM13 cuda compatible with the current PyTorch installation.
RuntimeError: CUDA error: no kernel image is available for execution on the device

System Configuration:

  • OS: Redhat 8.6
  • GPU Machine: H100 80GB HBM13
  • CUDA Version: 12.0
  • PyTorch Version: 2.0.1(python package)
    value for print(torch.version.cuda) = 11.7
  • GPU Driver Version: Nvidia-smi 525.60.13 driver version : 525.60.13

I’m running out of ideas on how to resolve this issue. If anyone has encountered a similar problem or has suggestions on what I might try next, I’d greatly appreciate your guidance.

Thank you in advance for your help.

Your Hopper GPU needs CUDA 11.8, so either install the latest stable release with 11.8 or the nightly binaries with CUDA 11.8 or 12.1.