UMM-JEPA proposes masked next-scale prediction for unified multimodal pretraining
A new arXiv paper, UMM-JEPA, introduces generative pretraining via masked next-scale representation prediction, aiming to build unified multimodal models that support both visual understanding and generation.
Read the full story at arXiv cs.CV ↗
Timeline · 1 report
- 2026-11-17 05:00 · arXiv cs.CV
UMM-JEPA: Generative Pretraining via Masked Next-Scale Representation Prediction for Unified Multimodal Models