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

[Feature] Support the training of QDTrack algorithm in MOT #433

Merged
merged 7 commits into from
Mar 6, 2022
Merged

[Feature] Support the training of QDTrack algorithm in MOT #433

merged 7 commits into from
Mar 6, 2022

Conversation

Seerkfang
Copy link
Collaborator

No description provided.

@Seerkfang Seerkfang changed the title Add QDTrack configs. [Add] QuasiD-Dense Tracking training model Feb 17, 2022
@codecov
Copy link

codecov bot commented Feb 17, 2022

Codecov Report

Merging #433 (9828804) into master (c8ddba2) will increase coverage by 0.61%.
The diff coverage is 89.41%.

Impacted file tree graph

@@            Coverage Diff             @@
##           master     #433      +/-   ##
==========================================
+ Coverage   72.51%   73.12%   +0.61%     
==========================================
  Files         121      125       +4     
  Lines        7054     7218     +164     
  Branches     1332     1364      +32     
==========================================
+ Hits         5115     5278     +163     
+ Misses       1538     1519      -19     
- Partials      401      421      +20     
Flag Coverage Δ
unittests 73.08% <89.41%> (+0.66%) ⬆️

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

Impacted Files Coverage Δ
mmtrack/models/mot/qdtrack.py 75.00% <75.00%> (ø)
...track/models/track_heads/quasi_dense_embed_head.py 87.30% <87.30%> (ø)
...track/models/track_heads/quasi_dense_track_head.py 94.28% <94.28%> (ø)
mmtrack/models/losses/__init__.py 100.00% <100.00%> (ø)
mmtrack/models/losses/l2_loss.py 76.19% <100.00%> (+55.55%) ⬆️
...track/models/losses/multipos_cross_entropy_loss.py 100.00% <100.00%> (ø)
mmtrack/models/mot/__init__.py 100.00% <100.00%> (ø)
mmtrack/models/mot/base.py 52.47% <100.00%> (ø)
mmtrack/models/track_heads/__init__.py 100.00% <100.00%> (ø)
mmtrack/models/track_heads/roi_embed_head.py 85.34% <100.00%> (ø)
... and 3 more

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update c8ddba2...9828804. Read the comment docs.

@Seerkfang Seerkfang requested review from GT9505 and removed request for GT9505 February 17, 2022 07:01
Copy link
Collaborator

@GT9505 GT9505 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

miss all docstrings

mmtrack/models/losses/l2_loss.py Show resolved Hide resolved
configs/mot/qdtrack/qdtrack_faster_rcnn_r50_fpn.py Outdated Show resolved Hide resolved
mmtrack/models/losses/multipos_cross_entropy_loss.py Outdated Show resolved Hide resolved
mmtrack/models/losses/multipos_cross_entropy_loss.py Outdated Show resolved Hide resolved
mmtrack/models/mot/qdtrack.py Outdated Show resolved Hide resolved
mmtrack/models/mot/qdtrack.py Outdated Show resolved Hide resolved
mmtrack/models/mot/qdtrack.py Outdated Show resolved Hide resolved
mmtrack/models/mot/qdtrack.py Outdated Show resolved Hide resolved
]
return dict(det_bboxes=det_bboxes, track_bboxes=track_bboxes)

def show_result(self,
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

use the show_result in BaseMultiObjectTracker


return losses

def simple_test(self, img, img_metas, rescale=False):
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

delete the function in the PR, since the tracker is not in the PR

mmtrack/models/mot/qdtrack.py Outdated Show resolved Hide resolved
mmtrack/models/track_heads/quasi_dense_embed_head.py Outdated Show resolved Hide resolved
mmtrack/models/track_heads/quasi_dense_embed_head.py Outdated Show resolved Hide resolved
mmtrack/models/track_heads/quasi_dense_embed_head.py Outdated Show resolved Hide resolved
"""bool: whether the track head contains a `embed_head`"""
return hasattr(self, 'embed_head') and self.embed_head is not None

def init_weights(self, *args, **kwargs):
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

please use init_cfg to init the weights of modules

mmtrack/models/track_heads/quasi_dense_track_head.py Outdated Show resolved Hide resolved
@Seerkfang Seerkfang requested a review from GT9505 March 2, 2022 13:00
Copy link
Collaborator

@GT9505 GT9505 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You should have been careful about the docstring.

mmtrack/models/losses/multipos_cross_entropy_loss.py Outdated Show resolved Hide resolved
mmtrack/models/losses/multipos_cross_entropy_loss.py Outdated Show resolved Hide resolved
mmtrack/models/mot/qdtrack.py Show resolved Hide resolved
mmtrack/models/mot/qdtrack.py Outdated Show resolved Hide resolved
tests/test_models/test_forward/test_qdtrack_forward.py Outdated Show resolved Hide resolved
mmtrack/models/mot/qdtrack.py Outdated Show resolved Hide resolved
mmtrack/models/track_heads/quasi_dense_track_head.py Outdated Show resolved Hide resolved
mmtrack/models/track_heads/quasi_dense_track_head.py Outdated Show resolved Hide resolved
@Seerkfang Seerkfang requested a review from GT9505 March 3, 2022 05:24
@@ -0,0 +1,100 @@
_base_ = './qdtrack_faster_rcnn_r50_fpn.py'
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

These three configs still need to be refactored.
Please discuss it with me.

@Seerkfang Seerkfang requested a review from GT9505 March 5, 2022 12:06
@GT9505 GT9505 changed the title [Add] QuasiD-Dense Tracking training model [Feature] Support the training of QDTrack algorithm in MOT Mar 6, 2022
@GT9505 GT9505 merged commit 8143901 into open-mmlab:master Mar 6, 2022
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.

None yet

2 participants