DB-GPT/docs/getting_started/install/llm/vllm/vllm.md

vLLM

vLLM is a fast and easy-to-use library for LLM inference and serving.

Running vLLM

Installing Dependencies

vLLM is an optional dependency in DB-GPT, and you can manually install it using the following command:

pip install -e ".[vllm]"

Modifying the Configuration File

Next, you can directly modify your .env file to enable vllm.

LLM_MODEL=vicuna-13b-v1.5
MODEL_TYPE=vllm

You can view the models supported by vLLM here

Then you can run it according to Run.