Dockerfile

FROM python:3.10-slim

# Set working directory
WORKDIR /app

# Set environment variables for cache directories
ENV HF_HOME=/tmp/.cache/huggingface
ENV TRANSFORMERS_CACHE=/tmp/.cache/huggingface/transformers
ENV TORCH_HOME=/tmp/.cache/torch
ENV PYTHONPATH=/app


# Install OpenMP
RUN apt-get update && apt-get install -y libomp-dev

# Install git
RUN apt-get update && apt-get install -y --no-install-recommends git

# Install compilers
RUN apt update && apt install -y gcc g++ clang clang-tools cmake

# Create cache directories with proper permissions
RUN mkdir -p /tmp/.cache/huggingface/transformers && \
    mkdir -p /tmp/.cache/torch && \
    chmod -R 777 /tmp/.cache

# Updating pip
RUN pip install --upgrade pip

# Install vllm cpu
ENV VLLM_TARGET_DEVICE=cpu
#ENV PIP_EXTRA_INDEX_URL=https://download.pytorch.org/whl/cpu
#RUN pip install git+https://github.com/vllm-project/vllm.git

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

# Copy application code
COPY . .

# Make sure app files are readable
RUN chmod -R 755 /app

# Expose port
EXPOSE 7860

# Run the application
CMD ["python", "app.py"]