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

We’re on a journey to advance and democratize artificial intelligence through open source and open science.