FROM runpod/pytorch:1.0.2-cu1281-torch280-ubuntu2404

ENV PYTHONUNBUFFERED=1 \
    PATH="/opt/runpod/venv/bin:${PATH}"

WORKDIR /opt/runpod

COPY templates/llm-inference/workspace/serve.sh /opt/runpod/bin/serve.sh

RUN chmod +x /opt/runpod/bin/serve.sh \
    && python3 -m venv /opt/runpod/venv \
    && . /opt/runpod/venv/bin/activate \
    && pip install --upgrade pip \
    && pip install vllm --extra-index-url https://download.pytorch.org/whl/cu128
