LensVLM: Compressing long context as images, expanding only relevant pages

60 points

by victormustar

8 hours ago

7 comments