search

PKU-YuanGroup / MoE-LLaVA

Mixture-of-Experts for Large Vision-Language Models
https://arxiv.org/abs/2401.15947
Apache License 2.0
1.9k stars 121 forks source link

[Question] Adding to the dataset. #58

Open arthurwolf opened 6 months ago

arthurwolf commented 6 months ago

Question

I'd like Llava to become better at some tasks I need it to do (in the next version, I guess).

(the tasks are related to reading/understanding comic book pages/panels, and analyzing their different elements)

Is it possible to propose additions to the dataset, for them to then be used in the next round of training?

If so, what is the correct repository and procedure to do this?

Any help would be greatly appreciated.

Thank you.