GenEval
A compositional evaluation benchmark for text-to-image generation that examines a model's understanding of the structured semantics of prompts through fine-grained subtasks such as object count, color, spatial relations, and attribute binding.
Timeline (last 90 days)
On the GenEval benchmark, LARC achieves a 7-percentage-point improvement over BAGEL, reaching state-of-the-art performance.
GenEval is a commonly used compositional evaluation benchmark in the text-to-image field, which tests a model's understanding of the structured semantics of prompts through fine-grained subtasks such as object count, color, spatial relations, and attribute binding.
LARC's improvement on GenEval indicates that the gains from continuous latent reasoning are mainly reflected in dimensions requiring precise semantic alignment such as compositional fidelity, rather than merely overall image aesthetics.
On the FLUX model, stacking SFG on top of CFG improves the GenEval score from 0.6661 to 0.7479.
All Facts (4)
On the GenEval benchmark, LARC achieves a 7-percentage-point improvement over BAGEL, reaching state-of-the-art performance.
50%UnverifiedGenEval is a commonly used compositional evaluation benchmark in the text-to-image field, which tests a model's understanding of the structured semantics of prompts through fine-grained subtasks such as object count, color, spatial relations, and attribute binding.
50%UnverifiedLARC's improvement on GenEval indicates that the gains from continuous latent reasoning are mainly reflected in dimensions requiring precise semantic alignment such as compositional fidelity, rather than merely overall image aesthetics.
50%UnverifiedOn the FLUX model, stacking SFG on top of CFG improves the GenEval score from 0.6661 to 0.7479.
50%