Dockerfile

# Use the NVIDIA CUDA image with CUDNN and development tools
FROM nvidia/cuda:12.3.2-cudnn9-devel-ubuntu22.04

# Install Python and pip
RUN apt-get update && \
    apt-get install -y  python3-dev python3-pip && \
    apt-get clean && \
    rm -rf /var/lib/apt/lists/*

# Install git
RUN apt-get update && apt-get install -y git

# Create a user with UID 1000
RUN useradd -m -u 1000 user
USER user
ENV PATH="/home/user/.local/bin:$PATH"

# Set the working directory
WORKDIR /app

# Copy the requirements file and install dependencies
COPY --chown=user ./requirements.txt requirements.txt
RUN pip3 install --upgrade pip
RUN pip3 install packaging
RUN pip3 install --no-cache-dir --upgrade -r requirements.txt


# Install numpy version compatible with flash_attn
RUN pip3 install numpy==1.23.4

# Install flash_attn
RUN pip3 install flash_attn

# Copy the rest of the application
COPY --chown=user . /app

# Command to run the application
CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]