vLLM Docker Container Image

vLLM is a fast and easy-to-use library for LLM inference and serving. This container image runs the OpenAI API server of vLLM.

The image is only for TPU and CPU inference. For GPU inference, please use the upstream image from vLLM.

Image URLs:

• substratusai/vllm (Docker Hub)
• ghcr.io/substratusai/vllm (GitHub Container Registry)

This image only publishes the TPU and CPU images for vLLM:

• substratusai/vllm:main-tpu
• substratusai/vllm:main-cpu

There are also tags available such as v0.6.3-tpu and v0.6.3-cpu.

Please use the upstream GPU image from vLLM directly:

vllm/vllm-openai:latest

Support the project by adding a star! ❤️

Join us on Discord:

Quickstart

Deploy Mistral 7B Instruct using Docker:

docker run -d -p 8080:8080 --gpus=all \
  substratusai/vllm \
  --model=mistralai/Mistral-7B-Instruct-v0.1

Building

docker build -t ghcr.io/substratusai/vllm .