Steering Visual Generation in Unified Multimodal Models with Understanding Supervision

Open paper

Benchmark	Model	Rank	Results
image-editing-on-gedit-bench-en	BAGEL-7B (UNO)	–	Overall: 7.17
text-to-image-generation-on-dpg-bench	BAGEL-7B (UNO)	–	DPG-Bench Score: 86.12