Finetuning on 16 Tesla K80 GPUs on EC2 Instance (p2.16xlarge) · Issue #747 – GitHub

Uncategorized

I am trying to finetune Vincuna-7B on 16 Tesla K80(12 GB) GPUs. I am getting Runtime: Expected is_sm90 || is_sm8x || is_sm75 to be true, …
View full source