Dynamic Pattern Alignment Learning for Pretraining Lightweight Human-Centric Vision Models

Re-assign community