Original model: Not directly usable with vLLM for faster inference.
Eren-Senoglu/llava-med-v1.5-mistral-7b-hf model: Good, but has bug currently
This version: Fully converted to the Hugging Face format and allows direct use of vLLM.

Thanks to this PR under Eren-Senoglu/llava-med-v1.5-mistral-7b-hf repo, I create this repo to have the fixed version.

All the contribution should go to Eren-Senoglu and xk-huang

Usage

vllm serve wnkh/llava-med-v1.5-mistral-7b-hf

vllm serve wnkh/llava-med-v1.5-mistral-7b-hf --revision aeedceb