Add uint8 bicubic support to ResizeV2 #7668
Merged
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
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/vision/7668
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit 8e6351b: This comment was automatically generated by Dr. CI and updates every 15 minutes. |
NicolasHug
commented
Jun 13, 2023
| ArgsKwargs((33, 26), interpolation=v2_transforms.InterpolationMode.BICUBIC, antialias=True), | ||
| ArgsKwargs((34, 25), interpolation=PIL.Image.BICUBIC, antialias=True), | ||
| ], | ||
| closeness_kwargs=dict(rtol=0, atol=21), |
There was a problem hiding this comment.
Had to pull these tests out so in order not to affect the atol=1 for the other tests.
vfdev-5
approved these changes
Jun 14, 2023
|
Thanks for the review and for the help with the tests @vfdev-5 . As a quick sanity check I evaluated the perf of some of our models that were trained with bicubic. Evaluation is done on PIL, float tensors and uint8 tensors (this PR). They're all extremely close. |
facebook-github-bot
pushed a commit
that referenced
this pull request
Jun 14, 2023
Summary: Co-authored-by: vfdev-5 <vfdev.5@gmail.com> Reviewed By: vmoens Differential Revision: D46724117 fbshipit-source-id: e47e4fbb2be67830fe56e22f1a806b8742d7652c
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Outdated suggestions cannot be applied.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.
Same as #7557 but for bicubic mode.
See pytorch/pytorch#103252 (comment) for AVX benchmarks. TL;DR: it's ~8X faster for tensors and ~4X for PIL.
Unlike bilinear mode, the bicubic uint8 path seems to be faster than the float path even on non-AVX archs:
cc @vfdev-5