Search Results for author:
Found 2 papers, 1 papers with code
Data-Free Knowledge Distillation Using Adversarially Perturbed OpenGL Shader Images
Knowledge distillation (KD) has been a popular and effective method for model compression.
Revisiting Batch Norm Initialization
Standard initialization of each BN in a network sets the affine transformation scale and shift to 1 and 0, respectively.
Cannot find the paper you are looking for? You can
Submit a new open access paper.
