[dartharva]

Does it support other visual-input-accepting language models? GPT-V is paywalled.

[Deverauxi]

The code isn’t very complicated. You could edit in any vision model you want:

https://github.com/microsoft/UFO/blob/main/ufo/llm/llm_call....

[dartharva]

I don't have the resources currently but would love to test something like LLaVA out for this, but I won't keep my hopes up.

[dartharva]

Update: no, it doesn't as of now.