Material classification on Masubuchi BN (T1), Graphene (T2), MoS2 (T3), WTe2 (T4)
92.68Accuracy (T1: Masubuchi BN)Joint Training
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Joint TrainingType=Ensemble, Backbone=ViT-B/16, Pre-trained=DeiT-Base [45]2025.08 | 92.68 | 92.04 | 90.91 | 92.82 | 92.11 | — | |
| Naive Fine-tuningTrained on=T1, Backbone=ViT-B/162025.08 | 91.46 | — | — | — | — | — | |
| CLIFFTrained on=T1, Backbone=ViT-B/162025.08 | 90.24 | — | — | — | — | — | |
| CLIFFTrained on=T2, Backbone=ViT-B/162025.08 | 86.59 | 79.33 | — | — | — | — | |
| L2PTrained on=T1, Backbone=ViT-B/162025.08 | 85.37 | — | — | — | — | — | |
| CLIFFTrained on=T3, Backbone=ViT-B/162025.08 | 64.63 | 77.7 | 79.87 | — | — | — | |
| L2PTrained on=T2, Backbone=ViT-B/162025.08 | 60.98 | 82.34 | — | — | — | — | |
| L2PTrained on=T3, Backbone=ViT-B/162025.08 | 57.32 | 52.88 | 87.66 | — | — | — | |
| CLIFFTrained on=T4, Backbone=ViT-B/162025.08 | 56.1 | 44.79 | 44.16 | 82.78 | 56.96 | 34.8 | |
| L2PTrained on=T4, Backbone=ViT-B/162025.08 | 53.66 | 21.23 | 1.3 | 71.77 | 36.99 | 59.73 | |
| Naive Fine-tuningTrained on=T2, Backbone=ViT-B/162025.08 | 10.98 | 86.09 | — | — | — | — | |
| Naive Fine-tuningTrained on=T3, Backbone=ViT-B/162025.08 | 8.54 | 22.23 | 83.77 | — | — | — | |
| Naive Fine-tuningTrained on=T4, Backbone=ViT-B/162025.08 | 4.88 | 3.2 | 0.65 | 62.68 | 17.85 | 84.2 |