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

[Improvement] Refactor Swin-Transformer #800

Merged
merged 13 commits into from
Sep 29, 2021

Conversation

xvjiarui
Copy link
Collaborator

Motivation

Refactor Swin-Transformer related code as open-mmlab/mmdetection#5748

BC-breaking (Optional)

We no longer support directly loading official weight here, instead, we will use the conversion script in #783

@codecov
Copy link

codecov bot commented Aug 21, 2021

Codecov Report

Merging #800 (dba93b0) into master (2aa632e) will increase coverage by 0.05%.
The diff coverage is 83.88%.

Impacted file tree graph

@@            Coverage Diff             @@
##           master     #800      +/-   ##
==========================================
+ Coverage   89.62%   89.67%   +0.05%     
==========================================
  Files         113      114       +1     
  Lines        6263     6431     +168     
  Branches      989     1007      +18     
==========================================
+ Hits         5613     5767     +154     
- Misses        452      462      +10     
- Partials      198      202       +4     
Flag Coverage Δ
unittests 89.67% <83.88%> (+0.05%) ⬆️

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

Impacted Files Coverage Δ
mmseg/models/backbones/swin.py 84.02% <75.00%> (-3.43%) ⬇️
mmseg/models/utils/embed.py 90.43% <90.81%> (+8.85%) ⬆️
mmseg/models/backbones/mit.py 91.33% <100.00%> (-0.20%) ⬇️
mmseg/models/backbones/vit.py 87.11% <100.00%> (ø)
mmseg/models/backbones/__init__.py 100.00% <0.00%> (ø)
mmseg/models/backbones/bisenetv1.py 100.00% <0.00%> (ø)

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 2aa632e...dba93b0. Read the comment docs.

@Junjun2016
Copy link
Collaborator

resolve some comments.

mmseg/models/utils/embed.py Outdated Show resolved Hide resolved
mmseg/models/utils/embed.py Outdated Show resolved Hide resolved
mmseg/models/utils/embed.py Outdated Show resolved Hide resolved
mmseg/models/utils/embed.py Outdated Show resolved Hide resolved
mmseg/models/utils/embed.py Outdated Show resolved Hide resolved
mmseg/models/utils/embed.py Outdated Show resolved Hide resolved
Copy link
Collaborator

@Junjun2016 Junjun2016 left a comment

Choose a reason for hiding this comment

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

LGTM, it seems that the code coverage should be improved.

@xvjiarui
Copy link
Collaborator Author

Coverage improved. The rest should come from the loading checkpoint part.

Copy link
Collaborator

@Junjun2016 Junjun2016 left a comment

Choose a reason for hiding this comment

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

LGTM and I will re-run the benchmark to double-check correctness.

@Junjun2016 Junjun2016 merged commit 85227b4 into open-mmlab:master Sep 29, 2021
bowenroom pushed a commit to bowenroom/mmsegmentation that referenced this pull request Feb 25, 2022
* [Improvement] Refactor Swin-Transformer

* fixed swin test

* update patch emebd, add more tests

* fixed test

* remove pretrain_style

* fixed padding

* resolve coments

* use mmcv 2tuple

* refactor init_cfg

Co-authored-by: Junjun2016 <hejunjun@sjtu.edu.cn>
wjkim81 pushed a commit to wjkim81/mmsegmentation that referenced this pull request Dec 3, 2023
* add LiteHRNet into README.md and README_CN.md
* minor fix to webcam demo and docs
sibozhang pushed a commit to sibozhang/mmsegmentation that referenced this pull request Mar 22, 2024
…b#800)

* first commit

* update changelog

* add verify in onnx unittest

* change unittest

Co-authored-by: dreamerlin <528557675@qq.com>
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