feat: Multimodal LLMs?

[Trickster85]

trafficstars

Feature request

Is it possible to rum multimodal LLMs like Qwen VL or LLaVa 1.5 using openllm?

Motivation

No response

Other

No response

[parano]

You can use BentoML for those models, which allow user to build custom, multi-modal APIs.

It's possible to support those models in OpenLLM in the future depending on community demand.