Human preference prediction on ImageReward 371 prompts (test)
39.62Recall @1ImageReward
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| ImageRewardBackbone=BLIP (ViT-L), Training Set Size=8k2023.04 | 39.62 | 63.07 | 90.84 | 49.06 | 70.89 | 88.95 | |
| Aesthetic Score2023.04 | 30.73 | 53.91 | 75.74 | 32.08 | 54.45 | 76.55 | |
| BLIP Score2023.04 | 30.73 | 50.67 | 77.63 | 33.42 | 56.33 | 80.59 | |
| CLIP Score2023.04 | 27.22 | 48.52 | 78.17 | 29.65 | 51.75 | 76.82 |