Distilling Self-Supervised Vision Transformers for Weakly-Supervised Few-Shot Classification & Segmentation | Read Paper on Bytez