FROM python:3.9 RUN useradd -m -u 1000 user USER user ENV PATH="/home/user/.local/bin:$PATH" ENV CMAKE_ARGS="-DGGML_OPENMP=off" ENV FORCE_CMAKE=1 WORKDIR /app # Téléchargement direct et robuste du modèle (2.4 Go) pendant la construction de l'image Docker RUN pip install --no-cache-dir --upgrade pip && \ wget -O llama-3.1-8B-Q2_K.gguf "https://huggingface.co/bartowski/Meta-Llama-3.1-8B-Instruct-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-Q2_K.gguf" COPY --chown=user ./requirements.txt requirements.txt RUN pip install --no-cache-dir --upgrade -r requirements.txt COPY --chown=user . /app CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"] RUN pip install --no-cache-dir --upgrade pip && \ wget -O llama-3.1-8B-Q2_K.gguf "https://huggingface.co/bartowski/Meta-Llama-3.1-8B-Instruct-GGUF/resolve/main/Meta-Llama-3.1-8B-Instruct-Q2_K.gguf" # Forcer le telechargement 1