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

(huggingface.co)

28 points | by victormustar  4 hours ago

2 comments