Dockerfile

FROM ubuntu:22.04

ENV MODEL_NAME="llava-1.6-mistral-7b-gguf"
ENV DEFAULT_MODEL_FILE="llava-v1.6-mistral-7b.Q3_K_XS.gguf"
ENV MODEL_USER="cjpais"
ENV DEFAULT_MODEL_BRANCH="main"
ENV DEFAULT_CLIP_MODEL_FILE="mmproj-model-f16.gguf"
ENV MODEL_URL="https://huggingface.co/${MODEL_USER}/${MODEL_NAME}/resolve/${DEFAULT_MODEL_BRANCH}/${DEFAULT_MODEL_FILE}"
ENV CLIP_MODEL_URL="https://huggingface.co/${MODEL_USER}/${MODEL_NAME}/resolve/${DEFAULT_MODEL_BRANCH}/${DEFAULT_CLIP_MODEL_FILE}"

ENV DEBIAN_FRONTEND=noninteractive

RUN apt update && \
    apt install --no-install-recommends -y build-essential python3 python3-pip wget curl git && \
    apt clean && rm -rf /var/lib/apt/lists/*

WORKDIR /app

COPY requirements.txt ./

RUN python3 -m pip install --upgrade pip && \
    pip install -r requirements.txt

# Downloading the models
RUN echo ${MODEL_URL} && \
    wget -O /app/${DEFAULT_MODEL_FILE} ${MODEL_URL} && \
    echo ${CLIP_MODEL_URL} && \
    wget -O /app/${DEFAULT_CLIP_MODEL_FILE} ${CLIP_MODEL_URL}

# Creating a non-root user
RUN useradd -m -u 1000 user && \
    mkdir -p /home/user/app && \
    mv /app/${DEFAULT_MODEL_FILE} /home/user/app && \
    mv /app/${DEFAULT_CLIP_MODEL_FILE} /home/user/app && \
    chown -R user:user /home/user/app

USER user
ENV HOME=/home/user \
    PATH=/home/user/.local/bin:$PATH

WORKDIR $HOME/app

# Copying the rest of your application
COPY --chown=user . .

RUN ls -al

EXPOSE 8000

CMD ["python3", "-m", "llama_cpp.server", "--model", "/home/user/app/llava-v1.6-mistral-7b.Q3_K_XS.gguf", "--clip_model_path", "/home/user/app/mmproj-model-f16.gguf", "--chat_format", "llava-1-5"]