Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

DPG-Bench

Benchmarks

Task NameDataset NameSOTA ResultTrend
Text-to-Image GenerationDPG-Bench
Overall Score88.63
265
Text-to-Image GenerationDPG-Bench
DPG Score88.9
131
Text-to-Image GenerationDPG-Bench (test)
Global Fidelity94.31
58
Text-to-image generationDPG-Bench
Average Score88.79
51
Text-to-Image GenerationDPG-Bench
DPG Score89
36
Dense prompt followingDPG-Bench v1.0 (test)
Entity Score93.08
20
Image GenerationDPG-Bench 130
Score88.3
15
Dense Prompt AlignmentDPG-Bench
Overall Score85.08
11
Text-to-3D GenerationDPG-Bench 1.0 (test)
Global Score81.82
7
Text-to-image alignmentDPG-Bench (test)
DPG77.13
6
Text-to-Image GenerationDPG-Bench zero-shot
DPG-Bench Score (Zero-Shot)70.1
5
Showing 11 of 11 rows