Fix PolynomialLR power type. #1440
Merged
+7
−6
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
@dotnet-policy-service agree company="HACARUS" |
|
regarding |
ds5678
approved these changes
Feb 8, 2025
alinpahontu2912
requested changes
Mar 13, 2025
It is a flaky test, we will try running again, don't worry |
hiyuh
force-pushed
the
fix-PolynomialLR-power-type
branch
from
802c1ef to
5b76bb7
Compare
|
hiyuh
force-pushed
the
fix-PolynomialLR-power-type
branch
from
5b76bb7 to
d1f9b3f
Compare
ozanMSFT
previously approved these changes
Mar 19, 2025
|
Hey @hiyuh, one of your PRs got merged, thanks for the contribution! Can you fix the Releasenotes conflict so we can merge this one too ? |
added 2 commits
March 19, 2025 21:06
torch.optim.lr_scheduler.PolynomialLR power was typed int, but should be double. Non-integer power is widely used for common training recipe. E.g. torchvision's pre-trained semantic segmentation models uses PolynomialLR as main LR scheduler with power = 0.9; https://github.com/pytorch/vision/blob/main/references/segmentation/train.py#L201 See also https://pytorch.org/docs/stable/generated/torch.optim.lr_scheduler.PolynomialLR.html
hiyuh
force-pushed
the
fix-PolynomialLR-power-type
branch
from
d1f9b3f to
79739db
Compare
|
@alinpahontu2912 |
alinpahontu2912
approved these changes
Mar 19, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
torch.optim.lr_scheduler.PolynomialLRpowerwas typedint, but should bedouble.Non-integer
poweris widely used for common training recipe.E.g.
torchvision's pre-trained semantic segmentation models usesPolynomialLRas main LR scheduler withpower = 0.9; https://github.com/pytorch/vision/blob/main/references/segmentation/train.py#L201See also https://pytorch.org/docs/stable/generated/torch.optim.lr_scheduler.PolynomialLR.html