Visplay Self Evolving Vlms

Enable VLMs to self-improve from unlabeled images via dual-role framework—questioner generates challenging visual questions while reasoner answers them, trained jointly with GRPO using difficulty and diversity rewards.

adu2021 Updated 7 repo stars

File contents

adu2021/skillxiv commit 9347ad45ac

Frequently asked questions

npx get-skill add adu2021/visplay-self-evolving-vlms