{"message":"unknown or invalid runtime name: nvidia"}

#26
by iholy19 - opened

Hello,
I am using GPU - 2 x NVIDIA Tesla V100 for inference in AzureML. I built the docker image successfully. But I get this error and it does not mention the code line number. I don't understand what I am missing. Please help me fix this.

ERROR:
Failed to execute command group with error API queried with a bad parameter: {"message":"unknown or invalid runtime name: nvidia"}

NVIDIA org

Hi @iholy19 ,

We haven't tried inference in AzureML. We recommend running the example python code in our docker environment.

Please let me know if you have further questions.

Thanks.
Best,
Boxin

Okat, Thank you.

I cloned the repo and I am trying to load the model. The repo is around 330GB. Could you please tell me how much of vRAM is needed?

managed to run in 4 bit quantization using 47gb vRAM, and 8bit quantization using 95gb vRAM

Sign up or log in to comment