Language Compositionality on SugarCrepe 1.0 (test)
88.1Recall@1 (Replace, Object)CLIP-PGS
Evaluation results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| CLIP-PGSMasking Ratio=0.3, Image Encoder=ViT-B/162025.03 | 88.1 | 76 | 67.9 | 64.1 | 66.5 | 74.2 | 69.9 | 75.5 | 70.8 | 67.9 | |
| E-CLIPImage Encoder=ViT-B/162025.03 | 86.9 | 73.5 | 60.2 | 59.4 | 63.4 | 73.3 | 66.8 | 73.2 | 68.4 | 60.2 | |
| A-CLIPImage Encoder=ViT-B/162025.03 | 86.6 | 75.5 | 63.2 | 52.4 | 63.1 | 71.6 | 66.8 | 71.6 | 68.4 | 63.2 | |
| CLIP-PGSMasking Ratio=0.3, Image Encoder=ViT-S/162025.03 | 86.6 | 77.8 | 63.9 | 58.8 | 63.4 | 74.9 | 68.4 | 73.4 | 70.2 | 63.9 | |
| CLIP-PGSMasking Ratio=0.5, Image Encoder=ViT-B/162025.03 | 86 | 77 | 64.6 | 63.3 | 65.5 | 77.3 | 69.8 | 75.5 | 70.8 | 64.6 | |
| CLIPImage Encoder=ViT-B/162025.03 | 85.8 | 79.2 | 64.5 | 61.8 | 58.7 | 74.2 | 68.4 | 73.7 | 68.8 | 64.5 | |
| CLIP-PGSMasking Ratio=0.5, Image Encoder=ViT-S/162025.03 | 84.9 | 76.5 | 65.4 | 60.4 | 65 | 73.6 | 69.9 | 73 | 70.5 | 65.4 | |
| FLIPImage Encoder=ViT-B/162025.03 | 84.1 | 75.9 | 66 | 60.2 | 61.6 | 71.7 | 63.2 | 72 | 66.9 | 66 |