Configuring Data Augmentations to Reduce Variance Shift in Positional Embedding of Vision Transformers

ViT
Re-assign community