Dockerfile

# Base image
FROM ghcr.io/ggerganov/llama.cpp:full-cuda

ENV DEBIAN_FRONTEND=noninteractive

# Update and install necessary dependencies
RUN apt update && \
    apt install --no-install-recommends -y \
        build-essential \
        python3 \
        python3-pip \
        wget \
        curl \
        git \
        cmake \
        zlib1g-dev \
        libblas-dev && \
    apt clean && \
    rm -rf /var/lib/apt/lists/*
    
# Setting up CUDA environment variables (this may not be necessary since you're using the official nvidia/cuda image, but it's good to be explicit)
ENV PATH="/usr/local/cuda/bin:$PATH" \
    LD_LIBRARY_PATH="/usr/local/cuda/lib64:$LD_LIBRARY_PATH" \
    CUDA_HOME="/usr/local/cuda"

WORKDIR /app

# Download ggml and mmproj models from HuggingFace
RUN wget https://huggingface.co/cjpais/llava-1.6-mistral-7b-gguf/resolve/main/llava-v1.6-mistral-7b.Q5_K_M.gguf && \
    wget https://huggingface.co/cjpais/llava-1.6-mistral-7b-gguf/resolve/main/mmproj-model-f16.gguf 

# Clone and build llava-server with CUDA support
RUN ls -al
RUN make LLAMA_CUBLAS=1

# Expose the port
EXPOSE 8080

# Start the llava-server with models
CMD ["--server", "--model", "llava-v1.6-mistral-7b.Q5_K_M.gguf", "--mmproj", "mmproj-model-f16.gguf", "--threads", "6", "--host", "0.0.0.0", "-ngl", "33"]