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.

Sign up for GitHub

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

Jump to bottom

[Refactor] Using MMCls backbones #233

Merged
merged 7 commits into from
Mar 21, 2022
Merged

[Refactor] Using MMCls backbones #233

merged 7 commits into from
Mar 21, 2022

Conversation

fangyixiao18
Copy link
Collaborator

Thanks for your contribution and we appreciate it a lot. The following instructions would make your pull request more healthy and more easily get feedback. If you do not understand some items, don't worry, just make the pull request and seek help from maintainers.

Motivation

Using backbones from mmcls,

Modification

modify resnet.py
modify resnext.py

BC-breaking (Optional)

Does the modification introduce changes that break the backward compatibility of the downstream repositories?
If so, please describe how it breaks the compatibility and how the downstream projects should modify their code to keep compatibility with this PR.

Use cases (Optional)

If this PR introduces a new feature, it is better to list some use cases here and update the documentation.

Checklist

Before PR:

  • Pre-commit or other linting tools are used to fix the potential lint issues.
  • Bug fixes are fully covered by unit tests, the case that causes the bug should be added in the unit tests.
  • The modification is covered by complete unit tests. If not, please add more unit test to ensure the correctness.
  • The documentation has been modified accordingly, like docstring or example tutorials.

After PR:

  • If the modification has potential influence on downstream or other related projects, this PR should be tested with those projects, like MMDet or MMSeg.
  • CLA has been signed and all committers have signed the CLA in this PR.

ZwwWayne
ZwwWayne reviewed Mar 16, 2022
View reviewed changes
mmselfsup/models/backbones/resnext.py

return nn.Sequential(*layers)


@BACKBONES.register_module()
Copy link

Choose a reason for hiding this comment

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

Why do we still need ResNeXt? Can we also import mmcls.ResNeXt?

Copy link
Collaborator Author

Choose a reason for hiding this comment

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

Why do we still need ResNeXt? Can we also import mmcls.ResNeXt?

As we are using forward function from our own ResNet, which is different from forward function in mmcls.

Copy link
Collaborator

Choose a reason for hiding this comment

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

Why do we still need ResNeXt? Can we also import mmcls.ResNeXt?

MMCls does not output the feature map from the first conv layer, which we will use for downstream evaluation. That's why we rewrite the forward function.

Copy link

Choose a reason for hiding this comment

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

  1. We should add comments in the code to explain the difference and the reason.
  2. We can give feedback to mmcls, e.g., create a feature request. Invite @mzr1996 to take a look.

@YuanLiuuuuuu YuanLiuuuuuu merged commit 6dbed8e into open-mmlab:dev_v0.8.0 Mar 21, 2022
@ZwwWayne ZwwWayne mentioned this pull request Mar 22, 2022
Closed
YuanLiuuuuuu pushed a commit that referenced this pull request Mar 31, 2022
@fangyixiao18 @YuanLiuuuuuu
[Refactor] Using MMCls backbones (#233)
d02c1aa 
* [Refactor] using backbones from MMCls

* [Refactor] modify the unit test

* [Fix] modify default setting of out_indices

* [Docs] fix lint

* [Refactor] modify super init

* [Refactore] remove res_layer.py

* using mmcv PatchEmbed
Jiahao000 added a commit that referenced this pull request Mar 31, 2022
@fangyixiao18 @YuanLiuuuuuu
@Jiahao000 @liming-ai @soonera
Bump version to v0.8.0 (#269)
df907e5 
* [Fix]: Fix mmcls upgrade bug (#235)

* [Feature]: Add multi machine dist_train (#232)

* [Feature]: Add multi machine dist_train

* [Fix]: Change bash to sh

* [Fix]: Fix missing sh suffix

* [Refactor]: Change bash to sh

* [Refactor] Add unit test (#234)

* [Refactor] add unit test

* update workflow

* update

* [Fix] fix lint

* update test

* refactor moco and densecl unit test

* fix lint

* add unit test

* update unit test

* remove modification

* [Feature]: Add MAE metafile (#238)

* [Feature]: Add MAE metafile

* [Fix]: Fix lint

* [Fix]: Change LARS to AdamW in the metafile of MAE

* [Fix] fix codecov bug (#241)

* [Fix] fix codecov bug

* update comment

* [Refactor] Using MMCls backbones (#233)

* [Refactor] using backbones from MMCls

* [Refactor] modify the unit test

* [Fix] modify default setting of out_indices

* [Docs] fix lint

* [Refactor] modify super init

* [Refactore] remove res_layer.py

* using mmcv PatchEmbed

* [Fix]: Fix outdated problem (#249)

* [Fix]: Fix outdated problem

* [Fix]: Update MoCov3 bibtex

* [Fix]: Use abs path in README

* [Fix]: Reformat MAE bibtex

* [Fix]: Reformat MoCov3 bibtex

* [Feature] Resume from the latest checkpoint automatically. (#245)

* [Feature] Resume from the latest checkpoint automatically.

* fix windows path problem

* fix lint

* add code reference

* [Docs] add docstring for ResNet and ResNeXt (#252)

* [Feature] support KNN benchmark (#243)

* [Feature] support KNN benchmark

* [Fix] add docstring and multi-machine testing

* [Fix] fix lint

* [Fix] change args format and check init_cfg

* [Docs] add benchmark tutorial

* [Docs] add benchmark results

* [Feature]: SimMIM supported (#239)

* [Feature]: SimMIM Pretrain

* [Feature]: Add mix precision and 16x128 config

* [Fix]: Fix config import bug

* [Fix]: Fix config bug

* [Feature]: Simim Finetune

* [Fix]: Log every 100

* [Fix]: Fix eval problem

* [Feature]: Add docstring for simmim

* [Refactor]: Merge layer wise lr decay to Default constructor

* [Fix]:Fix simmim evaluation bug

* [Fix]: Change model to be compatible to latest version of mmcls

* [Fix]: Fix lint

* [Fix]: Rewrite forward_train for classification cls

* [Feature]: Add UT

* [Fix]: Fix lint

* [Feature]: Add 32 gpus training for simmim ft

* [Fix]: Rename mmcls classifier wrapper

* [Fix]: Add docstring to SimMIMNeck

* [Feature]: Generate docstring for the forward function of simmim encoder

* [Fix]: Rewrite the class docstring for constructor

* [Fix]: Fix lint

* [Fix]: Fix UT

* [Fix]: Reformat config

* [Fix]: Add img resolution

* [Feature]: Add readme and metafile

* [Fix]: Fix typo in README.md

* [Fix]: Change BlackMaskGen to BlockwiseMaskGenerator

* [Fix]: Change the name of SwinForSimMIM

* [Fix]: Delete irrelevant files

* [Feature]: Create extra transformerfinetuneconstructor

* [Fix]: Fix lint

* [Fix]: Update SimMIM README

* [Fix]: Change SimMIMPretrainHead to SimMIMHead

* [Fix]: Fix the docstring of ft constructor

* [Fix]: Fix UT

* [Fix]: Recover deletion

Co-authored-by: Your <you@example.com>

* [Fix] add seed to distributed sampler (#250)

* [Fix] add seed to distributed sampler

* fix lint

* [Feature] Add ImageNet21k (#225)

* solve memory leak by limited implementation

* fix lint problem

Co-authored-by: liming <liming.ai@bytedance.com>

* [Refactor] change args format to '--a-b' (#253)

* [Refactor] change args format to `--a-b`

* modify tsne script

* modify 'sh' files

* modify getting_started.md

* modify getting_started.md

* [Fix] fix 'mkdir' error in prepare_voc07_cls.sh (#261)

* [Fix] fix positional parameter error (#260)

* [Fix] fix command errors in benchmarks tutorial (#263)

* [Docs] add brief installation steps in README.md (#265)

* [Docs] add colab tutorial (#247)

* [Docs] add colab tutorial

* fix lint

* modify the colab tutorial, using API to train the model

* modify the description

* remove #

* modify the command

* [Docs] translate 6_benchmarks.md into Chinese (#262)

* [Docs] translate 6_benchmarks.md into Chinese

* Update 6_benchmarks.md

change 基准 to 基准评测

* Update 6_benchmarks.md

(1)  Add Chinese translation of  ‘1 folder for ImageNet nearest-neighbor classification task’
(2) 数据预准备 -> 数据准备

* [Docs] remove install scripts in README (#267)

* [Docs] Update version information in dev branch (#268)

* update version to v0.8.0

* fix lint

* [Fix]: Install the latest mmcls

* [Fix]: Add SimMIM in RAEDME

Co-authored-by: Yuan Liu <30762564+YuanLiuuuuuu@users.noreply.github.com>
Co-authored-by: Jiahao Xie <52497952+Jiahao000@users.noreply.github.com>
Co-authored-by: Your <you@example.com>
Co-authored-by: Ming Li <73068772+mitming@users.noreply.github.com>
Co-authored-by: liming <liming.ai@bytedance.com>
Co-authored-by: RenQin <45731309+soonera@users.noreply.github.com>
Co-authored-by: YuanLiuuuuuu <3463423099@qq.com>
@fangyixiao18 fangyixiao18 deleted the mmcls-backbone branch April 1, 2022 06:18
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@ZwwWayne ZwwWayne ZwwWayne left review comments

@Jiahao000 Jiahao000 Jiahao000 approved these changes

@YuanLiuuuuuu YuanLiuuuuuu YuanLiuuuuuu approved these changes

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
4 participants
@fangyixiao18 @YuanLiuuuuuu @ZwwWayne @Jiahao000