Grey4sh
Results
12
comments of
Grey4sh
I encounterd the same problem when inference DeepSeekCoder-V2 on 8 A100s with the latest vLLM docker.
> Maybe use triton==2.2.0. For me, it solved this problem. It works for me.