Dockerfile

FROM --platform=amd64 nvcr.io/nvidia/cuda:12.1.0-devel-ubuntu22.04 as base

RUN apt update && \
    apt install -y python3-pip python3-packaging \
    git ninja-build && \
    pip3 install -U pip

# Tweak this list to reduce build time
# https://developer.nvidia.com/cuda-gpus
ENV TORCH_CUDA_ARCH_LIST "7.0;7.2;7.5;8.0;8.6;8.9;9.0"

RUN pip3 install "torch==2.1.1"

# This build is slow but NVIDIA does not provide binaries. Increase MAX_JOBS as needed.
# RUN pip3 install "git+https://github.com/stanford-futuredata/megablocks.git"
RUN pip3 install vllm
RUN pip3 install openai
RUN pip3 install "xformers==0.0.23" "transformers==4.36.0" "fschat[model_worker]==0.2.34"

RUN git clone https://github.com/NVIDIA/apex && \
    cd apex && git checkout 2386a912164b0c5cfcd8be7a2b890fbac5607c82 && \
    sed -i '/check_cuda_torch_binary_vs_bare_metal(CUDA_HOME)/d' setup.py && \
    python3 setup.py install --cpp_ext --cuda_ext

# Set up a new user named "user" with user ID 1000
RUN useradd -m -u 1000 user

# Switch to the "user" user
USER user

# Set home to the user's home directory
ENV HOME=/home/user \
	PATH=/home/user/.local/bin:$PATH

# Set the working directory to the user's home directory
WORKDIR $HOME/app

COPY --chown=user . $HOME/app

COPY --chown=user entrypoint.sh $HOME/app/

RUN ls -la $HOME/app/

RUN chmod +x $HOME/app/entrypoint.sh

ENTRYPOINT ["/home/user/app/entrypoint.sh"]