Fix naming of statistics in MeanAveragePrecision with custom max det thresholds
#2367
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
SkafteNicki
requested review from
Borda,
justusschock and
stancld
as code owners
SkafteNicki
changed the title
MeanAveragePrecision with custom max det thresholds
Codecov Report
Additional details and impacted files@@ Coverage Diff @@
## master #2367 +/- ##
========================================
- Coverage 69% 36% -33%
========================================
Files 305 305
Lines 17164 17164
========================================
- Hits 11836 6208 -5628
- Misses 5328 10956 +5628 |
Borda
approved these changes
Feb 12, 2024
r-remus
reviewed
Feb 12, 2024
| detection per image | ||
| - mar_{mdt[1]}: (:class:`~torch.Tensor`), mean average recall for `max_detection_thresholds[1]` (default 10) | ||
| detection per image | ||
| - mar_{mdt[1]}: (:class:`~torch.Tensor`), mean average recall for `max_detection_thresholds[2]` (default 100) |
There was a problem hiding this comment.
Is this a typo here? Should it maybe be
mar_{mdt[2]}: (:class:`~torch.Tensor`), mean average recall for `max_detection_thresholds[2]` (default 100)
mar_{mdt[1]}: (:class:`~torch.Tensor`), mean average recall for `max_detection_thresholds[2]` (default 100)
I realize this is merged already, but maybe this can be fixed boyscout-style in some future PR.
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Outdated suggestions cannot be applied.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.
What does this PR do?
Fixes #2360
Two fixes in this PR:
Beforehand we postulated that
faster-coco-evalsupported morelen(max_det_thresholds) > 3. However, looking at the code again, it is not true. Onlylen(max_det_thresholds)==3is actually supported, which we need to guard against. Here is the relevant code in both backends:https://github.com/MiXaiLL76/faster_coco_eval/blob/3475e078ec65e21f2ecf320b8493546357040a34/faster_coco_eval/core/cocoeval.py#L434-L442
https://github.com/cocodataset/cocoapi/blob/8c9bcc3cf640524c4c20a9c40e89cb6a2f2fa0e9/PythonAPI/pycocotools/cocoeval.py#L466-L468
(note how
self.params.maxDets[0],self.params.maxDets[1],self.params.maxDets[2]are hardcoded in both backends)Even when the user provides a custom list with length 3 like
[1,100,1000]the returned scores still saidmar_1,mar_10andmar_100(notmar_1000). This is just a simple renaming of the returned statistics.Before submitting
PR review
Anyone in the community is free to review the PR once the tests have passed.
If we didn't discuss your PR in Github issues there's a high chance it will not be merged.
Did you have fun?
Make sure you had fun coding 🙃
📚 Documentation preview 📚: https://torchmetrics--2367.org.readthedocs.build/en/2367/