Amir Zamir
Graph View
Backlinks
- 4M - Massively Multimodal Masked Modeling
- FlexTok - Resampling Images into 1D Token Sequences of Flexible Length
- MODUS - Decoder-Only Any-to-Any Modeling of Diverse Modalities
- Multimodality as Supervision - Self-Supervised Specialization to the Test Environment via Multimodality
- VideoFlexTok - Flexible-Length Coarse-to-Fine Video Tokenization