YellenYeason/yolov8-prune

Issues

• 请问如何进行稀疏训练呢

  #3 opened a year ago by ljqljq111
  1

• 当阈值大于最大权重时（keep_idx==0时）新的阈值是BN权重的最小值（15行），即该层不剪枝，这貌似不太合理哦

  #2 opened a year ago by aicmaodyu
  6

• 请问剪完后的训练越训越差是什么原因？

  #1 opened a year ago by Tomhardy13
  1