Image Classification on Food101 (Top-1 Accuracy, Expected Calibration Error)
84.7Top-1 AccuracyTPT
Evaluation results
| Method | Links | ||
|---|---|---|---|
| TPTPre-trained Backbone=CLIP Vit-B/16, Embedding dimension=1024-d2026.07 | 84.7 | 4.05 | |
| ARGTCA-DISCPre-trained Backbone=CLIP Vit-B/16, Embedding dimension=1024-d2026.07 | 83.85 | 2.06 | |
| ARGTCA-DIVPre-trained Backbone=CLIP Vit-B/16, Embedding dimension=1024-d2026.07 | 83.64 | 1.75 | |
| A-TPTPre-trained Backbone=CLIP Vit-B/16, Embedding dimension=1024-d2026.07 | 83.57 | 2.2 | |
| O-TPTPre-trained Backbone=CLIP Vit-B/16, Embedding dimension=1024-d2026.07 | 83.52 | 3.13 | |
| C-TPTPre-trained Backbone=CLIP Vit-B/16, Embedding dimension=1024-d2026.07 | 83.28 | 3.77 | |
| TCAPre-trained Backbone=CLIP Vit-B/16, Embedding dimension=1024-d2026.07 | 69.18 | 5.28 |