Dockerfile

# Use Python 3.10 as base image for better compatibility with ML libraries
FROM python:3.10-slim

# Set working directory
WORKDIR /app

# Install required system dependencies
RUN apt-get update && \
    apt-get install -y \
    git \
    wget \
    && apt-get clean \
    && rm -rf /var/lib/apt/lists/*

# Create and set permissions for directories
RUN mkdir -p /app/.cache/huggingface && \
    chmod 777 /app/.cache/huggingface && \
    mkdir -p /app/.git && \
    chmod 777 /app/.git

# Set environment variables
ENV TRANSFORMERS_CACHE=/app/.cache/huggingface/hub
ENV HF_HOME=/app/.cache/huggingface
ENV GIT_CONFIG_GLOBAL=/app/.git/config

# Copy requirements first to leverage Docker cache
COPY requirements.txt .

# Install Python dependencies
RUN pip install --no-cache-dir -r requirements.txt

# Create checkpoints directory with proper permissions
RUN mkdir -p /app/main/checkpoints && \
    chmod 777 /app/main/checkpoints

# Download model using litgpt command line with proper secret mounting
RUN --mount=type=secret,id=HF_TOKEN,mode=0444,required=true \
    export HF_TOKEN=$(cat /run/secrets/HF_TOKEN) && \
    echo "Starting model download..." && \
    litgpt download mistralai/Mistral-7B-Instruct-v0.3 \
        --access_token ${HF_TOKEN} \
        --checkpoint_dir /app/main/checkpoints || { echo "Download failed with status $?"; exit 1; }

# Copy the rest of the application
COPY . .

# Set environment variables for the application
ENV LLM_ENGINE_HOST=0.0.0.0
ENV LLM_ENGINE_PORT=7860
ENV MODEL_PATH=/app/main/checkpoints/mistralai/Mistral-7B-Instruct-v0.3

# Expose port 7860 for Hugging Face Spaces
EXPOSE 7860

# Command to run the application
CMD ["python", "main/main.py"]