Image Styling on Regular Text (test)
77.86Overall ScoreStandardized Reward-Weighted
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| Standardized Reward-WeightedModel Size=8B, Modality=Text, Training Method=SW2026.03 | 77.86 | 74.53 | 82.4 | 81.4 | 79.4 | 77 | 72.47 | |
| RModel Size=8B, Modality=Text, Training Method=R2026.03 | 77.62 | 73.33 | 81.8 | 81.53 | 78.53 | 76.6 | 73.8 | |
| Reward-WeightedModel Size=8B, Modality=Text, Training Method=RW2026.03 | 77.34 | 73.13 | 83 | 81.93 | 79.67 | 75.4 | 70.93 | |
| DPOModel Size=8B, Modality=Text, Training Method=D2026.03 | 75.85 | 72.13 | 79.93 | 78.8 | 77 | 74.2 | 73 | |
| BaselineModel Size=8B, Modality=Text, Training Method=B2026.03 | 74.79 | 69.47 | 81.73 | 79 | 77.33 | 73.33 | 67.87 | |
| SModel Size=8B, Modality=Text, Training Method=S2026.03 | 74.23 | 68.8 | 81.13 | 78.87 | 76.73 | 72.93 | 67.27 | |
| GPT-4oModel Size=N/A, Modality=Text, Training Method=G4o, Mode=Zero-shot2026.03 | 73.7 | 73.2 | 77.33 | 76.87 | 75.8 | 75.27 | 75.8 | |
| Edit-OnlyModel Size=8B, Modality=Text, Training Method=E2026.03 | 71.24 | — | 77.53 | — | — | 67 | — |