Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Fix] enable CocoMetric to use bbox score in pred_instances #2145

Merged
merged 3 commits into from
Apr 4, 2023

Conversation

Ben-Louis
Copy link
Collaborator

@Ben-Louis Ben-Louis commented Apr 4, 2023

Motivation

Adjust the CocoMetric to utilize the predicted bounding box score as the instance score derived from pred_instances.bbox_scores.

Modification

  1. Update CocoMetric to incorporate the predicted bounding box score as the instance score from pred_instances.bbox_scores.
  2. Modify the testing configuration for yolox-pose and revise its accuracy.
  3. Allow the inferencer to filter instances using predicted bounding box scores. Since the bounding box score threshold has been lowered to 0.001 from 0.2 in modification # 2, the model will yield low-quality instances. Therefore, a subsequent filtering step is required.

BC-breaking (Optional)

Use cases (Optional)

Checklist

Before PR:

  • I have read and followed the workflow indicated in the CONTRIBUTING.md to create this PR.
  • Pre-commit or linting tools indicated in CONTRIBUTING.md are used to fix the potential lint issues.
  • Bug fixes are covered by unit tests, the case that causes the bug should be added in the unit tests.
  • New functionalities are covered by complete unit tests. If not, please add more unit tests to ensure correctness.
  • The documentation has been modified accordingly, including docstring or example tutorials.

After PR:

  • CLA has been signed and all committers have signed the CLA in this PR.

@codecov
Copy link

codecov bot commented Apr 4, 2023

Codecov Report

Patch coverage: 48.38% and project coverage change: +0.02 🎉

Comparison is base (896e9d5) 82.25% compared to head (e3e10ea) 82.27%.

❗ Current head e3e10ea differs from pull request most recent head c09fc5f. Consider uploading reports for the commit c09fc5f to get more accurate results

Additional details and impacted files
@@             Coverage Diff             @@
##           dev-1.x    #2145      +/-   ##
===========================================
+ Coverage    82.25%   82.27%   +0.02%     
===========================================
  Files          228      228              
  Lines        13387    13393       +6     
  Branches      2268     2269       +1     
===========================================
+ Hits         11011    11019       +8     
+ Misses        1862     1855       -7     
- Partials       514      519       +5     
Flag Coverage Δ
unittests 82.27% <48.38%> (+0.02%) ⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files Coverage Δ
mmpose/evaluation/metrics/coco_metric.py 95.00% <0.00%> (-0.96%) ⬇️
mmpose/apis/inferencers/pose2d_inferencer.py 70.58% <33.33%> (-2.22%) ⬇️
mmpose/visualization/local_visualizer.py 62.94% <51.92%> (+0.13%) ⬆️
mmpose/apis/inferencers/mmpose_inferencer.py 75.60% <100.00%> (+0.30%) ⬆️

... and 2 files with indirect coverage changes

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.

@Ben-Louis Ben-Louis force-pushed the fix-coco-metric branch 2 times, most recently from 5dc765c to c09fc5f Compare April 4, 2023 11:28
@Tau-J Tau-J merged commit 37bb15c into open-mmlab:dev-1.x Apr 4, 2023
@Ben-Louis Ben-Louis deleted the fix-coco-metric branch April 26, 2023 18:00
shuheilocale pushed a commit to shuheilocale/mmpose that referenced this pull request May 6, 2023
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants