Custom
Benchmarks
TaskDatasetSOTA resultTrendResults
Custom
27.04PSNR
24
Custom (test)
71.4SSR
14
Custom English-to-Hebrew
30.09Win Rate (GPT-4o Judge)
11
Custom evaluation dataset
5.62Sync-C
9
Custom 200 samples (test)
0.328CLIP-T
9
Custom 120-image dataset Cycle Round 10 1.0 (test)
0.3LPIPS (Photograph-Creature)
8
Custom plushy (test)
37.34PSNR
8
Custom 27-video (test)
0.996Motion Smoothness
7
Custom 100 video samples (test)
18.127PSNR
6
Custom
0Failure Rate
5
Custom 200-sample (test)
0.298CLIP Similarity (T)
5
Custom (test)
18.11Mean Azimuth Error
4
Custom101
76.9Text Alignment
4
Custom V3 (test)
66.98Reference Consistency
4
Custom 5-scene 50 queries (test)
84Accuracy
3
Custom
98.2F1 Score
1
Custom
97.4Accuracy
1
Custom
95.9Accuracy
1