ViTKD: Feature-based Knowledge Distillation for Vision Transformers2024-06-170 PRE AI DOI:10.1109/cvprw63382.2024.00145原文链接原文求助分享收藏摘要 En