| # Base Python image | |
| FROM python:3.10-slim | |
| # Quiet + writable caches for HF on Spaces | |
| ENV PYTHONDONTWRITEBYTECODE=1 \ | |
| PYTHONUNBUFFERED=1 \ | |
| HF_HOME=/tmp/hf \ | |
| TRANSFORMERS_CACHE=/tmp/hf \ | |
| TOKENIZERS_PARALLELISM=false | |
| # Workdir | |
| WORKDIR /app | |
| # Upgrade pip and install deps first (better layer caching) | |
| COPY requirements.txt . | |
| RUN pip install --upgrade pip \ | |
| && pip install --no-cache-dir -r requirements.txt | |
| # Copy your LoRA export and app code | |
| COPY medbot_model /app/medbot_model | |
| COPY app.py inference.py ./ | |
| COPY templates ./templates | |
| # Port used by Spaces | |
| EXPOSE 8080 | |
| # Production server | |
| CMD ["gunicorn", "-w", "1", "-b", "0.0.0.0:8080", "app:app"] | |