Table 13.
Performance comparison of different fine-tuning strategies and MCTS on tomato disease datasets using Qwen3-VL-8B as the base model. (Part 2/2).
| MCTS | FT Strategy | Diseases | MCC | F1 | GM | Pre | Recall | ACC | Overall_ACC |
|---|---|---|---|---|---|---|---|---|---|
| ✗ | ST-R4 | B.S | 0.9173 | 0.9333 | 0.9750 | 0.8909 | 0.9800 | 0.9800 | 0.9320 |
| E.B | 0.8456 | 0.8667 | 0.8810 | 0.9750 | 0.7800 | 0.7800 | |||
| Healthy | 0.9877 | 0.9901 | 0.9975 | 0.9804 | 1.0000 | 1.0000 | |||
| L.B | 0.8554 | 0.8846 | 0.9398 | 0.8519 | 0.9200 | 0.9200 | |||
| Y.L.C.V | 0.9750 | 0.9800 | 0.9875 | 0.9800 | 0.9800 | 0.9800 | |||
| ✓ | ST-R4 | B.S | 0.9629 | 0.9703 | 0.9850 | 0.9608 | 0.9800 | 0.9800 | 0.9480 |
| E.B | 0.8598 | 0.8764 | 0.8832 | 1.0000 | 0.7800 | 0.7800 | |||
| Healthy | 1.0000 | 1.0000 | 1.0000 | 1.0000 | 1.0000 | 1.0000 | |||
| L.B | 0.8801 | 0.9009 | 0.9721 | 0.8197 | 1.0000 | 1.0000 | |||
| Y.L.C.V | 0.9875 | 0.9899 | 0.9899 | 1.0000 | 0.9800 | 0.9800 |