Dockerfile

ARG BASE_IMAGE=ghcr.io/huggingface/text-generation-inference:2.0.2

FROM python:3 as model
RUN pip install huggingface_hub

ARG MODEL_ID=HuggingFaceH4/zephyr-7b-beta
RUN huggingface-cli download ${MODEL_ID} --exclude *.bin --exclude *.pth --local-dir=/model --local-dir-use-symlinks=False

FROM ${BASE_IMAGE}

# Copy the downloaded model directory to the container
RUN mkdir -p /model
COPY --from=model /model /model