katsura-jp/pytorch-cosine-annealing-with-warmup

Issues

• method name of get_le() not consistent with torch.optim optimzers

  #17 opened 6 months ago by tanasr
  0

• Learning rate goes lower than specified min_lr

  #16 opened a year ago by lennarteing
  0

• ModuleNotFoundError: No module named 'cosine_annealing_warmup'

  #9 opened 2 years ago by Waterkin
  2

• Allow `max_lr` to be set per group

  #14 opened 2 years ago by katsura-jp
  0

• Warmup steps only apply on the first cycle

  #13 opened 2 years ago by katsura-jp
  0

• Additional Features

  #11 opened 2 years ago by SteveImmanuel
  1

• License?

  #10 opened 2 years ago by shreyaskamathkm
  1

• Is there a good reason why T_mult should be an integer

  #4 opened 4 years ago by JulienMaille
  2

• Weird gamma behavior

  #8 opened 3 years ago by dofuuz
  1

• Is there possibility to add verbose=True, to see increase/decrease in lr as training progresses.

  #7 opened 3 years ago by rkakash59
  1

• Citation to use this scheduler

  #5 opened 4 years ago by scaomath
  2

• base_lr relies on the lr of optimizer

  #3 opened 4 years ago by DietDietDiet
  0