Tony-Y/pytorch_warmup

Issues

• A BUG in BaseWarmup?

  #26 opened 3 months ago by Moon0316
  8

• Nice work!

  #25 opened 3 months ago by Suasy
  1

• Can the warmup_scheduler update the learning rate every epoch and not every batch?

  #17 opened 9 months ago by talrub
  2

• How to schedule LR with warmup on global_step initially, and then epoches after warmup?

  #16 opened a year ago by JohnHerry
  9

• About the learning rate in scheduler

  #15 opened 2 years ago by Arios42
  1

• My lr jumped from 0.01 to 0.0498 without any linear signs.

  #14 opened 2 years ago by lxt160980
  2

• UserWarning: Detected call of `lr_scheduler.step()` before `optimizer.step()`

  #13 opened 2 years ago by Angryrou
  2

• How to use in `pytorch-lightning`?

  #8 opened 2 years ago by sieu-n
  2

• Why did my learning rate drop from the initial lr

  #12 opened 2 years ago by sjchasel
  3

• Unexpected keyword argument `warmup_period`

  #11 opened 2 years ago by fbragman
  1

• License file is not included in sdist

  #9 opened 2 years ago by awvwgk
  1

• no attribute named dampening

  #7 opened 3 years ago by beegica
  1

• UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible.

  #5 opened 3 years ago by Mao-KU
  2

• difference of this library with hugging face

  #3 opened 3 years ago by brando90
  3

• Why is warmup better than RAdam?

  #2 opened 3 years ago by brando90
  3

• What is the decay rule of thumb?

  #4 opened 3 years ago by brando90
  3