Video Classification on HMDB-51 base-to-novel
77.1Base AccuracyVL prompting
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| VL promptingPre-trained=Kinetics-4002022.12 | 77.1 | 54.9 | 64.1 | |
| XCLIPPre-trained=Kinetics-4002022.12 | 75.8 | 52 | 61.7 | |
| TACOEvaluation Protocol=Base-to-novel2026.06 | 74.4 | 62.3 | 67.8 | |
| FROSTERText Augmentation=LLM-based2024.04 | 74.1 | 58 | 65.1 | |
| FROSTEREvaluation Protocol=Base-to-novel2026.06 | 74.1 | 58 | 65.1 | |
| ViFi-CLIPText Augmentation=Original2024.04 | 73.8 | 53.3 | 61.9 | |
| ViFi-CLIPEvaluation Protocol=Base-to-novel2026.06 | 73.8 | 53.3 | 61.9 | |
| Open-MeDeEvaluation Protocol=Base-to-novel2026.06 | 73.6 | 56.4 | 63.9 | |
| TC-CLIPText Augmentation=Original2024.04 | 73.3 | 54.1 | 62.2 | |
| TC-CLIPText Augmentation=LLM-based2024.04 | 73.3 | 59.1 | 65.5 | |
| A5Pre-trained=Kinetics-4002022.12 | 70.4 | 51.7 | 59.6 | |
| Open-VCLIPText Augmentation=Original2024.04 | 70.3 | 50.4 | 58.9 | |
| Open-VCLIPEvaluation Protocol=Base-to-novel2026.06 | 70.3 | 50.4 | 58.7 | |
| X-CLIPText Augmentation=Original2024.04 | 69.4 | 45.5 | 55 | |
| X-CLIPEvaluation Protocol=Base-to-novel2026.06 | 69.4 | 45.5 | 55 | |
| ActionCLIPText Augmentation=Original2024.04 | 69.1 | 37.3 | 48.5 | |
| ActionCLIPEvaluation Protocol=Base-to-novel2026.06 | 69.1 | 37.3 | 48.5 | |
| ActionCLIPPre-trained=Kinetics-4002022.12 | 69 | 57.2 | 62.6 | |
| ST-AdapterEvaluation Protocol=Base-to-novel2026.06 | 65.3 | 48.9 | 55.9 | |
| AIMEvaluation Protocol=Base-to-novel2026.06 | 64 | 51.6 | 57.1 | |
| Vanilla CLIPPre-trained=Kinetics-4002022.12 | 53.3 | 46.8 | 49.8 | |
| Vanilla CLIPText Augmentation=Original2024.04 | 53.3 | 46.8 | 49.8 | |
| CLIPEvaluation Protocol=Base-to-novel2026.06 | 53.3 | 46.8 | 49.8 | |
| A5Text Augmentation=Original2024.04 | 46.2 | 16 | 23.8 | |
| VPTEvaluation Protocol=Base-to-novel2026.06 | 46.2 | 16 | 23.8 |