@tduncan / Nvidia-Optimization-Stack-VLLM Mirror · read-only / commits / eb17212858
Fork 0

A high-throughput and memory-efficient inference and serving engine for LLMs

Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git

Synced Nov 17, 05:30 UTC

Update Dockerfile to support Mixtral (#2027)

Simon Mo committed Dec 11, 2023 at 11:59 UTC eb17212858dce5c73b8fe3e8c548861ea18f1f43
1 file changed +2 -2
Dockerfile
+2 -2
@@ -73,9 +73,9 @@ ENTRYPOINT ["python3", "-m", "vllm.entrypoints.api_server"]
73
74 # openai api server alternative
75 FROM vllm-base AS vllm-openai
76 -# install additional dependencies for openai api server
76 +# install additional dependencies for openai api server, and mixtral
77 RUN --mount=type=cache,target=/root/.cache/pip \
78 - pip install accelerate fschat
78 + pip install accelerate megablocks
79
80 COPY --from=build /workspace/vllm/*.so /workspace/vllm/
81 COPY vllm vllm