Classification on Moonstone Age (Downstream)
73.2AccuracyMG-MAE
Evaluation results
| Method | Links | |
|---|---|---|
| MG-MAEPre-training=Lunar, Evaluation Protocol=Finetune2026.07 | 73.2 | |
| MG-MAEPre-training=Lunar, Evaluation Protocol=Linear2026.07 | 69.8 | |
| Vanilla MAEPre-training=Lunar, Details=Per-channel tokenizers without modality grouping2026.07 | 63.5 | |
| ScratchPre-training=None2026.07 | 58.7 | |
| ImageNet ViT-BPre-training=ImageNet, Evaluation Protocol=Linear2026.07 | 50.4 |