Dockerfile

FROM python:3.11.9-slim-bullseye

USER root

RUN adduser --quiet --disabled-password \
    --home /home/conv_user \
    --shell /bin/bash conv_user
RUN adduser conv_user sudo

WORKDIR /srv
ENV PYTHONPATH="/srv"

RUN apt-get update && apt-get install -y git && rm -rf /var/lib/apt/lists/*

RUN apt-get -y update
RUN apt-get install -y --no-install-recommends \
    build-essential \
    python3-dev \
    libglib2.0-0 \
    libgl1-mesa-glx \
    libpq-dev \
    make
# insall latest vllm
RUN pip install vllm==0.7.2

# Install transformers from source, need specific revision for Qwen2.5 as 
# breaking changes in master
RUN pip install --upgrade git+https://github.com/huggingface/transformers.git@336dc69d63d56f232a183a3e7f52790429b871ef

ENTRYPOINT ["sh", "-c", "python -m vllm.entrypoints.openai.api_server --model convergence-ai/proxy-lite-3b --trust-remote-code --tokenizer-pool-size 10 --max_model_len 16384 --limit-mm-per-prompt image=1 --enable-auto-tool-choice --tool-call-parser hermes --port 7860"]