YOLOv3-model-pruning icon indicating copy to clipboard operation
YOLOv3-model-pruning copied to clipboard

situable prune rule

Open PiseyYou opened this issue 6 years ago • 3 comments

in step 2, you metion about the situable prune rule, is there any example to show? And what called suitable prune rule?

PiseyYou avatar Jul 29 '19 11:07 PiseyYou

@PiseyYou 其实基本上就是(1)避免剪掉所有channel的最高阈值(每个BN层的gamma的最大值的最小值即为阈值上限) (2)剪枝的时候会先对 bias 做处理,用其后的卷积层或者 BN 层来吸收这个参数,然后再将权重赋值给 compact model, 这样做的话就能保证 pruned_model (只对 weight 置 0 的模型)和 Compact_model 对相同的输入有相同的输出

violet17 avatar Jul 30 '19 03:07 violet17

@violet17 ,对的,太感谢你了,帮我回答了挺多问题的,我最近都没空回复issue

Lam1360 avatar Jul 31 '19 12:07 Lam1360

@Lam1360 不用客气,我也在学习,只是顺手复制了你别的回答过来 : )

violet17 avatar Aug 01 '19 05:08 violet17