gpu.Dockerfile

ARG CUDA_IMAGE="12.1.1-devel-ubuntu22.04"
FROM nvidia/cuda:${CUDA_IMAGE}
ENV MODEL=WizardLM-13B-V1.2

# Install the package
RUN apt-get update && apt-get upgrade -y \
    && apt-get install -y git build-essential \
    python3 python3-pip gcc wget \
    ocl-icd-opencl-dev opencl-headers clinfo \
    libclblast-dev libopenblas-dev \
    && mkdir -p /etc/OpenCL/vendors && echo "libnvidia-opencl.so.1" > /etc/OpenCL/vendors/nvidia.icd \
    && rm -rf /var/lib/apt/lists/*

ENV CUDA_DOCKER_ARCH=all
ENV LLAMA_CUBLAS=1

RUN python3 -m pip install --upgrade pip pytest cmake scikit-build setuptools fastapi uvicorn sse-starlette pydantic-settings requests

RUN CMAKE_ARGS="-DLLAMA_CUBLAS=on" FORCE_CMAKE=1 pip install llama-cpp-python

# Set a working directory for better clarity
WORKDIR /app

COPY ./start-llama2.sh /app/start-llama2.sh
COPY ./hug_model.py /app/hug_model.py

# Tell LLAMA_CPP that we want to offload layers to the GPU
ENV LLAMA_CPP_ARGS="--n_gpu_layers=43"

# Set environment variable for the host
ENV HOST=0.0.0.0

# Expose a port for the server
EXPOSE 8000

# Run the server start script
CMD ["/app/start-llama2.sh"]