Fine-grained Image Classification on FGVC-A
88.78Accuracy (Iteration 1)SelfSynthX
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| SelfSynthXtraining_mode=rejection sampling with detailed visual knowledge2025.02 | 88.78 | 90.91 | 91.42 | 91.99 | |
| NLtraining_mode=only train with labels2025.02 | 83.47 | 87.28 | 87.82 | 87.73 | |
| L+GEtraining_mode=train with labels and general explanations2025.02 | 72.13 | 79.87 | 82.45 | 82.69 | |
| Basetraining_mode=original model2025.02 | 3 | — | — | — |