Dockerfile

# syntax=docker/dockerfile:1

FROM python:3.10-slim

ENV PIP_NO_CACHE_DIR=1 \
    PYTHONDONTWRITEBYTECODE=1 \
    PYTHONUNBUFFERED=1 \
    HF_HUB_ENABLE_HF_TRANSFER=1 \
    HF_HOME=/data/hf \
    TRANSFORMERS_CACHE=/data/hf \
    DIFFUSERS_CACHE=/data/hf \
    HUGGINGFACE_HUB_CACHE=/data/hf

WORKDIR /app

# System deps
RUN apt-get update && apt-get install -y --no-install-recommends \
    git \
    gcc \
    libgl1 \
    libglib2.0-0 \
    libgthread-2.0-0 \
    libgomp1 \
    libgcc-s1 \
    && rm -rf /var/lib/apt/lists/*

# Copy code
COPY . /app

# Install deps
RUN python -m pip install --upgrade pip && \
    python -m pip install -r requirements.txt && \
    # Install CUDA-enabled PyTorch wheels (CUDA 12.1); required for GPU on Spaces
    python -m pip install --index-url https://download.pytorch.org/whl/cu121 torch torchvision torchaudio --upgrade --force-reinstall && \
    # Ensure xformers is not installed (optional perf lib; skip for compatibility)
    python -m pip uninstall -y xformers || true && \
    # Speed up and lower RAM during model load (avoid upgrading hub)
    python -m pip install --no-deps accelerate && \
    # Ensure compatible hub version providing required symbols
    python -m pip install --upgrade --force-reinstall "huggingface-hub==0.20.3" "diffusers==0.23.0"

EXPOSE 7860

ENV DATA_DIR=/data
RUN mkdir -p /data /data/hf && chmod -R 777 /data /data/hf
CMD ["python", "-m", "uvicorn", "server:app", "--host", "0.0.0.0", "--port", "7860"]