Search papers, labs, and topics across Lattice.
3
99
6
15
ReToken boosts vision-language model performance by over 20% on key benchmarks while fitting training and inference within a single GPU's memory constraints.
Vision-language models struggle to adapt plans based on visual input alone, revealing a critical gap in their ability to use what they see when things don't go as expected.
Forget task-specific models: Magma, a single foundation model, now outperforms them in both UI navigation and robotic manipulation by bridging verbal and action abilities.
