Fetching the paper…

Beyond Masking: Demystifying Token-Based Pre-Training for Vision Transformers · Around