Fine-grained Image Classification on Average (5 datasets) Macro-average (test)
89.3Base AccuracyToolFG
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| ToolFGTraining Strategy=per-dataset training, Evaluation Protocol=Zero-shot base-to-novel generalization2026.06 | 89.3 | 86.6 | 87.9 | |
| DiVE-kTraining Strategy=per-dataset training, Evaluation Protocol=Zero-shot base-to-novel generalization2026.06 | 80.8 | 78.8 | 79.8 | |
| PromptSRCTraining Strategy=per-dataset training, Evaluation Protocol=Zero-shot base-to-novel generalization2026.06 | 79.7 | 67.2 | 73 | |
| Fine-R1Training Strategy=per-dataset training, Evaluation Protocol=Zero-shot base-to-novel generalization2026.06 | 78.7 | 76.7 | 77.7 | |
| ViRFTTraining Strategy=per-dataset training, Evaluation Protocol=Zero-shot base-to-novel generalization2026.06 | 76.8 | 77.1 | 76.9 | |
| MaPLeTraining Strategy=per-dataset training, Evaluation Protocol=Zero-shot base-to-novel generalization2026.06 | 76.3 | 65.7 | 70.6 | |
| Qwen2.5-VL-7BTraining Strategy=per-dataset training, Evaluation Protocol=Zero-shot base-to-novel generalization2026.06 | 68.9 | 70.5 | 69.7 |