Dockerfile

# Use an official CUDA runtime as a parent image
FROM nvidia/cuda:12.1-cudnn8-runtime-ubuntu20.04

# Set the working directory in the container
WORKDIR /usr/src/app

# Install system dependencies
RUN apt-get update && apt-get install -y \
    ffmpeg \
    build-essential \
    cmake \
    git \
    && rm -rf /var/lib/apt/lists/*

# Install Python 3.12 and pip
RUN apt-get update && apt-get install -y python3.12 python3.12-venv python3.12-dev && \
    curl -sS https://bootstrap.pypa.io/get-pip.py | python3.12

# Set Python 3.12 as the default python
RUN update-alternatives --install /usr/bin/python python /usr/bin/python3.12 1

# Copy the current directory contents into the container at /usr/src/app
COPY . .

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

# Install Flash Attention 2 dependencies
RUN pip install packaging ninja
RUN pip install flash-attn --no-build-isolation

# Expose port 7860 for Gradio
EXPOSE 7860

# Define environment variable to avoid Python buffering
ENV PYTHONUNBUFFERED=1

# Set the entry point for the container to run any script
ENTRYPOINT ["python", "app.py"]