Regression on Moonstone Composition Downstream
0.924R2MG-MAE
Evaluation results
| Method | Links | |
|---|---|---|
| MG-MAEPre-training=Lunar, Evaluation Protocol=Linear2026.07 | 0.924 | |
| MG-MAEPre-training=Lunar, Evaluation Protocol=Finetune2026.07 | 0.908 | |
| Vanilla MAEPre-training=Lunar, Details=Per-channel tokenizers without modality grouping2026.07 | 0.871 | |
| ScratchPre-training=None2026.07 | 0.802 | |
| ImageNet ViT-BPre-training=ImageNet, Evaluation Protocol=Linear2026.07 | 0.631 |