Skip to content

[Fix] Add max_keep_ckpts for groupfree3d#765

Merged
ZwwWayne merged 1 commit intoopen-mmlab:masterfrom
yezhen17:groupfree3d
Jul 20, 2021
Merged

[Fix] Add max_keep_ckpts for groupfree3d#765
ZwwWayne merged 1 commit intoopen-mmlab:masterfrom
yezhen17:groupfree3d

Conversation

@yezhen17
Copy link
Copy Markdown
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

The size of a most simple groupfree3d ckpt is around 170MB, and there are by default 400 ckpts saved in one run, which will take about 70GB space. There is a high chance of disk space/quota exceed, which will terminate training.

Modification

I added max_keep_ckpts in groupfree3d configs.

BC-breaking (Optional)

Does the modification introduce changes that break the back-compatibility of the downstream repos?
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

  1. Pre-commit or other linting tools are used to fix the potential lint issues.
  2. The modification is covered by complete unit tests. If not, please add more unit test to ensure the correctness.
  3. If the modification has potential influence on downstream projects, this PR should be tested with downstream projects.
  4. The documentation has been modified accordingly, like docstring or example tutorials.

@yezhen17
Copy link
Copy Markdown
Collaborator Author

yezhen17 commented Jul 19, 2021

@hjin2902 , please take a look. I cannot request review from you. If you approve, you can also update #702 ?

@yezhen17 yezhen17 requested a review from ZwwWayne July 19, 2021 13:53
@codecov
Copy link
Copy Markdown

codecov bot commented Jul 19, 2021

Codecov Report

Merging #765 (8ff6ee9) into master (78f4562) will not change coverage.
The diff coverage is n/a.

Impacted file tree graph

@@           Coverage Diff           @@
##           master     #765   +/-   ##
=======================================
  Coverage   48.97%   48.97%           
=======================================
  Files         208      208           
  Lines       15852    15852           
  Branches     2532     2532           
=======================================
  Hits         7763     7763           
  Misses       7591     7591           
  Partials      498      498           
Flag Coverage Δ
unittests 48.97% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out 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 78f4562...8ff6ee9. Read the comment docs.

@hjin2902
Copy link
Copy Markdown
Contributor

@hjin2902 , please take a look. I cannot request review from you. If you approve, you can also update #702 ?

Sorry for the late reply!
Yes, we can add config like this, but a potential problem is that the best checkpoint may not in the last 10 epoch. So maybe the value of max_keep_ckpts should be large enough.

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.

3 participants