Dockerfile

FROM vllm/vllm-openai:latest

# Expose your desired port
EXPOSE 7860

# Environment variables for host/port
ENV VLLM_HOST=0.0.0.0
ENV VLLM_PORT=7860
ENV VLLM_LOGGING_LEVEL=DEBUG

# Run vLLM with env-based host and port
CMD ["sh", "-c", "vllm serve --model unsloth/llama-2-7b-bnb-4bit --host 0.0.0.0 --port 7860"]