sanaka87/BAGEL-RecA
Any-to-Any
•
Updated
•
133
•
25
RecA
Unlocking the Massive Zero-shot Potential in Unified Multimodal Models through Self-supervised Learning!
-
-
sanaka87/Harmon-0.5B-RecA
Text-to-Image • Updated • 15 • 4 -
sanaka87/Harmon-1.5B-RecA
Any-to-Any • Updated • 16 • 2 -
sanaka87/Show-o-RecA
Text-to-Image • Updated • 13 • 3 -
sanaka87/Show-o-512x512-RecA
Any-to-Any • Updated • 11 • 2 -
sanaka87/Harmon-1.5B-RecA-plus
Text-to-Image • Updated • 18 • 3 -
sanaka87/OpenUni-RecA
Any-to-Any • Updated • 16 • 1 -
Reconstruction Alignment Improves Unified Multimodal Models
Paper • 2509.07295 • Published • 39