openai/CLIP
1
0
Fork 0
mirror of https://github.com/openai/CLIP.git synced 2026-11-05 01:37:26 +00:00
Code Issues 274 Projects Packages Wiki Activity

[GH-ISSUE #107] LR Schedule Hyperparameters #98

New issue
Closed
opened 2026-10-31 21:11:05 +00:00 by bsakaguchi · 1 comment
bsakaguchi commented 2026-10-31 21:11:05 +00:00
Owner
Copy link

Originally created by @Zasder3 on GitHub (May 17, 2021).
Original GitHub issue: https://github.com/openai/CLIP/issues/107

Hello, I'm trying to replicate your LR scheduling as described within the paper and have a couple of questions:

  1. Is your scheduler the one with or without restarts, i.e. is it torch.optim.lr_scheduler.CosineAnnealingWarmRestarts?
  2. When you specify warmup iterations, is that the period (T_0) of the cosine function or is that something different altogether?

Thank you!

Originally created by @Zasder3 on GitHub (May 17, 2021). Original GitHub issue: https://github.com/openai/CLIP/issues/107 Hello, I'm trying to replicate your LR scheduling as described within the paper and have a couple of questions: 1. Is your scheduler the one with or without restarts, i.e. is it `torch.optim.lr_scheduler.CosineAnnealingWarmRestarts`? 2. When you specify warmup iterations, is that the period (`T_0`) of the cosine function or is that something different altogether? Thank you!
bsakaguchi commented 2026-10-31 21:11:06 +00:00
Author
Owner
Copy link

@jongwook commented on GitHub (Jun 30, 2021):

  1. The LR schedule didn't have restarts, the multiplier becomes (almost) 1.0 after warmup, and monotonically decreases 0.0 drawing a cosine curve over 32 epochs.
  2. A linear warmup was done over the first 2,000 iterations and is not dependent of the period of the cosine function.
<!-- gh-comment-id:871276202 --> @jongwook commented on GitHub (Jun 30, 2021): 1. The LR schedule didn't have restarts, the multiplier becomes (almost) 1.0 after warmup, and monotonically decreases 0.0 drawing a cosine curve over 32 epochs. 2. A linear warmup was done over the first 2,000 iterations and is not dependent of the period of the cosine function.
Sign in to join this conversation.
No Branch/Tag specified
Branches Tags
main
codex/pin-github-workflow-refs-20260325-132838
jongwook/issue-396
jongwook-patch-1
No results found.
Labels
Clear labels   
pull-request
Mirrored from GitHub Pull Request

No labels
pull-request
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
openai/CLIP#98
No description provided.