UniCom: Unified Multimodal Modeling via Compressed Continuous Semantic Representations

Open paper

Benchmark	Model	Rank	Results
image-generation-on-wise	UniCom	–	Overall: 0.58
text-to-image-generation-on-dpg-bench	UniCom	–	DPG-Bench Score: 85.92
text-to-image-generation-on-geneval	UniCom	–	Overall: 0.87