ColorMAE: Exploring data-independent masking strategies in Masked AutoEncoders

Benchmark Model Rank Results
image-classification-on-imagenetColorMAE-Green-ViTB-1600#380Top 1 Accuracy: 83.8%
semantic-segmentation-on-ade20kColorMAE-Green-ViTB-1600#132Validation mIoU: 49.3