Dockerfile

# Multi-stage build for MVDiff
FROM nvidia/cuda:11.8.0-cudnn8-runtime-ubuntu22.04 AS base

# Set environment variables
ENV PYTHONUNBUFFERED=1 \
    PYTHONDONTWRITEBYTECODE=1 \
    DEBIAN_FRONTEND=noninteractive \
    CUDA_VISIBLE_DEVICES=0

# Install system dependencies
RUN apt-get update && apt-get install -y \
    python3.10 \
    python3-pip \
    git \
    wget \
    libgl1-mesa-glx \
    libglib2.0-0 \
    libsm6 \
    libxext6 \
    libxrender-dev \
    libgomp1 \
    libglib2.0-0 \
    && rm -rf /var/lib/apt/lists/*

# Set working directory
WORKDIR /app

# Copy requirements first for better caching
COPY requirements.txt .

# Install Python dependencies
RUN pip3 install --no-cache-dir --upgrade pip setuptools wheel && \
    pip3 install --no-cache-dir torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118 && \
    pip3 install --no-cache-dir -r requirements.txt

# Copy application code
COPY . .

# Create necessary directories
RUN mkdir -p checkpoints data outputs logs

# Download default model weights (optional)
# RUN python3 scripts/download_models.py --model mvdiff-base

# Expose ports
EXPOSE 7860 6006

# Health check
HEALTHCHECK --interval=30s --timeout=10s --start-period=5s --retries=3 \
    CMD curl -f http://localhost:7860/health || exit 1

# Default command
CMD ["python3", "app.py", "--port", "7860"]

# Development stage
FROM base AS dev

# Install development dependencies
RUN pip3 install --no-cache-dir \
    pytest \
    pytest-cov \
    black \
    flake8 \
    mypy \
    jupyterlab \
    ipdb

# Expose additional ports for development
EXPOSE 8888

# Production stage
FROM base AS prod

# Add non-root user
RUN useradd -m -u 1000 mvdiff && \
    chown -R mvdiff:mvdiff /app

USER mvdiff

# Set production environment
ENV PYTHONOPTIMIZE=1

# Run with gunicorn for production
CMD ["gunicorn", "--bind", "0.0.0.0:7860", "--workers", "1", "--threads", "4", "app:app"]