M.I.M.I.R - Multi-agent Intelligent Memory & Insight Repository

# Build llama.cpp server for ARM64 (Apple Silicon) with Nomic Embed Multimodal FROM ubuntu:22.04 AS builder # Install build dependencies RUN apt-get update && apt-get install -y \ build-essential \ cmake \ git \ curl \ libcurl4-openssl-dev \ && rm -rf /var/lib/apt/lists/* # Clone llama.cpp repository WORKDIR /build RUN git clone https://github.com/ggerganov/llama.cpp.git . # Build llama.cpp server with embeddings support RUN mkdir build && cd build && \ cmake .. \ -DLLAMA_CURL=ON \ -DLLAMA_BUILD_SERVER=ON \ -DBUILD_SHARED_LIBS=OFF \ -DCMAKE_BUILD_TYPE=Release && \ cmake --build . --config Release -j$(nproc) # Runtime stage FROM ubuntu:22.04 # Install runtime dependencies RUN apt-get update && apt-get install -y \ libcurl4 \ curl \ libgomp1 \ && rm -rf /var/lib/apt/lists/* # Copy built binaries COPY --from=builder /build/build/bin/llama-server /usr/local/bin/llama-server # Create models directory RUN mkdir -p /models # Copy Nomic Embed Multimodal model # Model will be downloaded by build script and placed here COPY docker/llama-cpp/models/nomic-embed-multimodal.gguf /models/nomic-embed-multimodal.gguf # Expose port EXPOSE 8080 # Health check HEALTHCHECK --interval=30s --timeout=10s --start-period=30s --retries=3 \ CMD curl -f http://localhost:8080/health || exit 1 # Set working directory WORKDIR /app # Default command with Nomic Embed Multimodal # Note: Adjust --ctx-size and embedding dimensions based on actual model specs ENTRYPOINT ["/usr/local/bin/llama-server"] CMD ["--host", "0.0.0.0", "--port", "8080", "--model", "/models/nomic-embed-multimodal.gguf", "--embeddings", "--pooling", "mean", "--ctx-size", "8192", "--parallel", "4"]