Visual Narrative Generation on VinaBench
72.6Non-Character Entity AlignmentDynaVieW
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| DynaVieWEvaluation Protocol=Supervised Fine-Tuning, Judge Model=GPT-4o2026.07 | 72.6 | 41.4 | 63.1 | 54.9 | 51.1 | 87.9 | 48.3 | 60.1 | 61 | |
| BAGELEvaluation Protocol=Supervised Fine-Tuning, Judge Model=GPT-4o2026.07 | 72.3 | 42 | 62.2 | 53.3 | 51.2 | 77 | 34.4 | 48.3 | 54.7 | |
| ARLDMEvaluation Protocol=Supervised Fine-Tuning, Judge Model=GPT-4o2026.07 | 70.1 | 40.3 | 62.1 | 46.2 | 46.1 | 80.6 | 31.3 | 32.9 | 50.6 | |
| Story2BoardEvaluation Protocol=Zero-Shot Generalization, Judge Model=GPT-4o2026.07 | 68.3 | 35.3 | 61.1 | 39.2 | 37.4 | 98.4 | 47.2 | 49 | 56.6 | |
| MM-InterleavedEvaluation Protocol=Supervised Fine-Tuning, Judge Model=GPT-4o2026.07 | 65.9 | 42.5 | 62.5 | 51.8 | 49.3 | 79.5 | 33.6 | 40.4 | 52.8 | |
| BAGELEvaluation Protocol=Zero-Shot Generalization, Judge Model=GPT-4o2026.07 | 62.2 | 31.5 | 52.8 | 59.9 | 49.2 | 82.4 | 37 | 67.5 | 56.7 | |
| DynaVieWEvaluation Protocol=Zero-Shot Generalization, Judge Model=GPT-4o2026.07 | 60 | 32.3 | 52.1 | 55.1 | 50.1 | 89.7 | 55 | 83.5 | 63 | |
| StoryGenEvaluation Protocol=Supervised Fine-Tuning, Judge Model=GPT-4o2026.07 | 56.6 | 38.8 | 41.2 | 31.3 | 41.6 | 57.6 | 15.8 | 34.1 | 38.9 | |
| Emu2Evaluation Protocol=Zero-Shot Generalization, Judge Model=GPT-4o2026.07 | 53.3 | 22 | 55.6 | 52.3 | 46.6 | 37.8 | 35.2 | 43.9 | 42.5 |