Refactoring ConvModule by removing act_cfg

[sungchul2]

Summary

This PR includes:

• Replace act_cfg with activation_callable
• Remove build_activation_layer

How to test

Checklist

• I have added unit tests to cover my changes.​
• I have added integration tests to cover my changes.​
• I have ran e2e tests and there is no issues.
• I have added the description of my changes into CHANGELOG in my target branch (e.g., CHANGELOG in develop).​
• I have updated the documentation in my target branch accordingly (e.g., documentation in develop).
• I have linked related issues.

License

• I submit my code changes under the same Apache License that covers the project.
  Feel free to contact the maintainers if that's a concern.
• I have updated the license header for each file (see an example below).

# Copyright (C) 2024 Intel Corporation
# SPDX-License-Identifier: Apache-2.0

[harimkang]

LGTM!

[eunwoosh]

Thanks for your work! I left some comments. Please take a look.

[harimkang]

I've canceled the action of this PR due to issues with the integration test at the moment, please update when this PR is merged. #3815

[kprokofi]

Thank you for the update. I have a general question.
Can't we pass and use already instantiated activation instead of activation class (partial)?

[codecov]

Codecov Report

Attention: Patch coverage is 96.89922% with 4 lines in your changes missing coverage. Please review.

Project coverage is 80.67%. Comparing base (df498b7) to head (ce0d48c).

Files	Patch %	Lines
src/otx/algo/segmentation/backbones/litehrnet.py	55.55%	4 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff            @@
##           develop    #3809   +/-   ##
========================================
  Coverage    80.66%   80.67%           
========================================
  Files          272      272           
  Lines        27477    27459   -18     
========================================
- Hits         22165    22153   -12     
+ Misses        5312     5306    -6     

Flag	Coverage Δ
py310	80.64% <96.89%> (+0.23%)	⬆️
py311	80.42% <96.89%> (-0.14%)	⬇️

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

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[sungchul2]

Thank you for the update. I have a general question. Can't we pass and use already instantiated activation instead of activation class (partial)?

I considered this before, but I haven't proceeded for now because original ConvModule's purpose is an instantiation using given dict config, such as conv_cfg, norm_cfg, or act_cfg.
I just updated this module maintaining this purpose, but I will try your suggestion if you or others (@harimkang, @eunwoosh, or any others 😊) think that using conv, norm, or activation instances as is is needed.
I think adding this is not hard work.

[eunwoosh]

Thank you for the update. I have a general question. Can't we pass and use already instantiated activation instead of activation class (partial)?

I considered this before, but I haven't proceeded for now because original ConvModule's purpose is an instantiation using given dict config, such as conv_cfg, norm_cfg, or act_cfg. I just updated this module maintaining this purpose, but I will try your suggestion if you or others (@harimkang, @eunwoosh, or any others 😊) think that using conv, norm, or activation instances as is is needed. I think adding this is not hard work.

I actually think this approach when reviewing norm refactoring but it also can be activation layer. I think current approach is ok but Kirill's opinion is also good.

[sungchul2]

Thank you for the update. I have a general question. Can't we pass and use already instantiated activation instead of activation class (partial)?

I'll try this targeting the next version as a new minor feature, conv and normalization can be also appiled.