Generic Personal Instance Retrieval on DeepFashion2 1.0 (test)
53.4mAPMeta-Personalized Vision-Language Model
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Meta-Personalized Vision-Language ModelBackbone=ViT-B/162023.06 | 53.4 | 77.7 | |
| Meta-Personalized Vision-Language ModelBackbone=ViT-B/322023.06 | 40 | 69.3 | |
| CLIP (visual)2023.06 | 20.6 | 43.7 | |
| CLIP (V+L)2023.06 | 20.5 | 42.8 | |
| CLIP (language)2023.06 | 8.3 | 16.9 | |
| Random2023.06 | 4.7 | 9.5 |