Dockerfile

# Use NVIDIA CUDA base image with Python support
FROM nvidia/cuda:12.1.1-runtime-ubuntu22.04

# Set working directory
WORKDIR /app

# Install Python and system dependencies
RUN apt-get update && apt-get install -y \
    python3.10 \
    python3-pip \
    python3.10-venv \
    git \
    curl \
    wget \
    && rm -rf /var/lib/apt/lists/*

# Make python3.10 the default
RUN update-alternatives --install /usr/bin/python python /usr/bin/python3.10 1

# Upgrade pip
RUN pip install --upgrade pip

# ✅ Install prebuilt llama-cpp-python CUDA wheel (cu121)
RUN pip install --no-cache-dir llama-cpp-python --extra-index-url https://abetlen.github.io/llama-cpp-python/whl/cu121

# Copy requirements.txt and install remaining dependencies
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

# Copy app code
COPY . .

# Create models directory
RUN mkdir -p models

# Expose Gradio or web port
EXPOSE 7860

# Gradio settings
ENV GRADIO_SERVER_NAME="0.0.0.0"
ENV GRADIO_SERVER_PORT=7860

# Start app
CMD ["python", "app.py"]