Dockerfile

FROM nvidia/cuda:12.0.0-cudnn8-devel-ubuntu22.04

# Set environment variables
ENV PATH="/usr/local/cuda/bin:$PATH"
ENV MODEL_NAME="falcon-40b-instruct-GGML"
ENV MODEL_FILE="falcon40b-instruct.ggmlv3.q4_K_S.bin"
ENV MODEL_URL="https://huggingface.co/TheBloke/${MODEL_NAME}/raw/ggmlv3/${MODEL_FILE}"

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/*

# Set the working directory in the container to /app
WORKDIR /app

# Install cmake
RUN apt-get install -y wget && \
    wget -qO- "https://cmake.org/files/v3.18/cmake-3.18.0-Linux-x86_64.tar.gz" | tar --strip-components=1 -xz -C /usr/local

# Copy the requirements.txt file into the container
COPY requirements.txt ./

# Install any needed packages specified in requirements.txt
RUN pip3 install --upgrade pip && \
    pip3 install -r requirements.txt

# Clone ctransformers and update submodule ggllm.cpp
RUN git clone --recursive https://github.com/marella/ctransformers.git && \
    cd ctransformers && \
    git submodule update --init models/submodules/ggllm.cpp && \
    cd models/submodules/ggllm.cpp && \
    git checkout master && \
    git pull

# Install ctransformers from source
RUN cd ctransformers && \
    CT_CUBLAS=1 FORCE_CMAKE=1 pip install .

# Download the model file
# RUN wget -O /app/${MODEL_FILE} ${MODEL_URL}

# Create user
RUN useradd -m -u 1000 user

# Create a directory for app and move the downloaded file there
# RUN mkdir -p /home/user/app && mv /app/${MODEL_FILE} /home/user/app

# Change the ownership of the copied file to user
# RUN chown user:user /home/user/app/${MODEL_FILE}

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

WORKDIR $HOME/app

# Now you can COPY the rest of your app
COPY --chown=user . .

RUN ls -al

# Make port available to the world outside this container
EXPOSE 7860

# Run uvicorn when the container launches
CMD ["python3", "demo.py", "--host", "0.0.0.0", "--port", "7860"]