running with flashmla on A100s

#1
by ehartford - opened
Cognitive Computations org

@v2ray is it possible you could make a docker image with your flashmla in it?

https://github.com/LagPixelLOL/vllm/tree/sm80_flashmla

Cognitive Computations org

I uploaded the wheel containing it to my org x2ray. I don't like to make Docker images because they are kind of messy.

Sign up or log in to comment