Reproduce “Teacher-Guided Student Self-Knowledge Distillation Using Diffusion Model” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Yu Wang, Chuanguang Yang, Zhulin An, Weilun Feng, Jiarui Zhao, Chengqing Yu, Libo Huang, Boyu Diao and 1 more
DSKD helps compress recognition models by refining student features through diffusion guided by the teacher’s classifier. It trains the student toward its own denoised features rather than directly matching teacher features. On ImageNet, a ResNet-18 student with a ResNet-34 teacher reaches 72.57% top-1 accuracy versus DiffKD’s 72.22% under B1 training; this comparison is specific to that setup.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.