Dockerfile

FROM nvidia/cuda:11.8.0-devel-ubuntu22.04

# Install necessary packages
RUN apt update && apt install -y \
    git \
    build-essential \
    libopenblas-dev \
    wget \
    python3-pip \
    nodejs \
    npm

# 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 and adjust PATH
ENV HOME=/home/user \
    PATH=/home/user/.local/bin:$PATH

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

# Copy the current directory contents into the container at $HOME/app setting the owner to the user
COPY --chown=user . $HOME/app

# Install aphrodite-engine and clone repository
RUN python3 -m pip install aphrodite-engine && \
    git clone https://github.com/lolcats399/aphrodite-engine

# Install Hugging Face libraries
RUN pip install huggingface-hub hf-transfer

# Set environment variable to enable hf-transfer
ENV HF_HUB_ENABLE_HF_TRANSFER=1

# Attempt to download goliath model - this line might cause delays due to model size.
RUN huggingface-cli download TheBloke/goliath-120b-gptq --local-dir $HOME/goliath-gptq --local-dir-use-symlinks False --cache-dir $HOME/cache

# Set environment variable for a bypass server (not sure how this is being used in your application)
ENV BYPASS_SERVER=http://localhost:8080

# Expose the port for the API server
EXPOSE 7860

# Run the aphrodite engine API server
CMD ["python3", "$HOME/app/aphrodite-engine/aphrodite/endpoints/kobold/api_server.py", "$ENGINE_ARGS", "--port", "7860", "-h", "--model", "$HOME/goliath-gptq"]