Optim-wip: Correct RandomScale transform
#922
Closed
Conversation
This file contains hidden or 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
Contributor
Author
|
Closing this PR as it is no longer needed |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.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.
The
RandomScaletransform should actually be usingF.interpolaterather thanF.affine_grid&F.grid_sample, so I've changed it. The old version ofRandomScalehowever seems to be useful enough to warrant keeping it, so I've renamed it toRandomScaleAffine.The
_rand_selectfunction is no longer being used by any transforms, so I've deleted it.Replaced default
RandomScaletransform with Interpolation based variant. Renamed the old variant toRandomScaleAffine.Added JIT support for both versions of
RandomScale. This required changing some of the type hints.Add distributions support to both versions of
RandomScale. Ludwig wanted this based off of his original PR.Added more tests to improve testing coverage.
Changed
NumSeqOrTensorTypehint toNumSeqOrTensorOrProbDistType.This PR contains code that was part of: #821, to make it easier to review.