Dockerfile

FROM debian:12-slim

ENV DEBIAN_FRONTEND=noninteractive
ENV VENV_PATH=/BitNet/venv
ENV PATH="$VENV_PATH/bin:$PATH"

# Install system dependencies
RUN apt-get update && apt-get install -y \
    wget \
    curl \
    git \
    cmake \
    build-essential \
    python3 \
    python3-pip \
    python3-venv \
    python-is-python3 \
    ca-certificates \
    && rm -rf /var/lib/apt/lists/*

# Clone BitNet
RUN git clone --recursive https://github.com/microsoft/BitNet.git

# Copy local files
COPY . /BitNet
WORKDIR /BitNet

# Create virtual environment
RUN python3 -m venv $VENV_PATH

# Upgrade pip and install dependencies inside venv
RUN python3 -m pip install --upgrade pip
RUN python3 -m pip install -r requirements.txt
RUN python3 -m pip install huggingface_hub

# Optional: Hugging Face token
ARG HF_TOKEN
RUN huggingface-cli login --token $HF_TOKEN || true

# Download model
RUN python3 -m huggingface_hub download HF1BitLLM/Llama3-8B-1.58-100B-tokens --local-dir models/Llama3-8B-1.58-100B-tokens

# Setup environment / quantization
RUN python3 -m setup_env -md models/Llama3-8B-1.58-100B-tokens -q i2_s

# Build C++ runtime
RUN mkdir -p build && cd build && cmake .. && make -j$(nproc)

# Default command
CMD ["python3", "app.py"]