Metric: Accuracy (higher is better)
| # | Model↕ | Accuracy▼ | Extra Data | Paper | Date↕ | Code |
|---|---|---|---|---|---|---|
| 1 | PATH (Partial FT) | 90.8 | Yes | HumanBench: Towards General Human-centric Percep... | 2023-03-10 | Code |
| 2 | Hulk(Finetune, ViT-L) | 88.97 | Yes | Hulk: A Universal Knowledge Translator for Human... | 2023-12-04 | Code |
| 3 | Hulk(Finetune, ViT-B) | 87.85 | Yes | Hulk: A Universal Knowledge Translator for Human... | 2023-12-04 | Code |
| 4 | C2T-Net | 87.2 | Yes | - | - | Code |
| 5 | SOLIDER | 86.38 | Yes | Beyond Appearance: a Semantic Controllable Self-... | 2023-03-30 | Code |
| 6 | UniHCP (finetune) | 86.18 | Yes | UniHCP: A Unified Model for Human-Centric Percep... | 2023-03-06 | Code |
| 7 | APTM | 80.17 | Yes | Towards Unified Text-based Person Retrieval: A L... | 2023-06-05 | Code |
| 8 | Label2Label | 79.23 | Yes | Label2Label: A Language Modeling Framework for M... | 2022-07-18 | Code |
| 9 | strongBaseline(ours) | 78.56 | No | Rethinking of Pedestrian Attribute Recognition: ... | 2020-05-25 | Code |