Zero Shot Video Question Answer On Intentqa
Métriques
Accuracy
Résultats
Résultats de performance de divers modèles sur ce benchmark
Tableau comparatif
| Nom du modèle | Accuracy |
|---|---|
| an-image-grid-can-be-worth-a-video-zero-shot | 65.3 |
| videotree-adaptive-tree-based-video | 66.9 |
| vidctx-context-aware-video-question-answering | 67.1 |
| a-simple-llm-framework-for-long-range-video | 64.0 |
| language-repository-for-long-video | 59.1 |
| self-chained-image-language-model-for-video-1 | 60.9 |
| too-many-frames-not-all-useful-efficient | 71.1 |
| enter-event-based-interpretable-reasoning-for | 71.5 |
| a-simple-llm-framework-for-long-range-video | 53.6 |
| mistral-7b | 50.4 |
| ts-llava-constructing-visual-tokens-through | 67.9 |
| slowfast-llava-a-strong-training-free | 60.1 |
| Modèle 13 | 20.0 |