Fetching the paper…

Uniform Masking: Enabling MAE Pre-training for Pyramid-based Vision Transformers with Locality · Around