AI & ML interests
None defined yet.
Recent Activity
View all activity
Papers
MODUS: Decoder-Only Any-to-Any Modeling of Diverse Modalities
(1D) Ordered Tokens Enable Efficient Test-Time Search
Multimodal datasets and models from https://tst-vision.epfl.ch
Multimodal models from https://4m.epfl.ch/
Omnidata surface normals and depth estimators
Autoregressive models trained on FlexTok.
Multimodal datasets and models from https://tst-vision.epfl.ch
Flexible 1D tokenizers and VAEs from https://flextok.epfl.ch/
Multimodal models from https://4m.epfl.ch/
Multimodal tokenizers from https://4m.epfl.ch/
Omnidata surface normals and depth estimators