Fix saved filename in ModelCheckpoint if it already exists by rohitgr7 · Pull Request #4861 · Lightning-AI/pytorch-lightning

rohitgr7 · 2020-11-25T22:26:55Z

What does this PR do?

Fixes the saved filename if it's already present. It will override the file if it's already going to be deleted. Previously if save_top_k=1 and filename is static for eg: some_file, it will create a file with the version for eg: filename=some_file-v2.ckpt if max_epoch > 1 but it should be some_file.ckpt only.

The test fails on master.

Before submitting

Was this discussed/approved via a Github issue? (no need for typos and docs improvements)
Did you read the contributor guideline, Pull Request section?
Did you make sure your PR does only one thing, instead of bundling different changes together? Otherwise, we ask you to create a separate PR for every change.
Did you make sure to update the documentation with your changes?
Did you write any new necessary tests?
Did you verify new and existing tests pass locally with your changes?
If you made a notable change (that affects users), did you update the CHANGELOG?

PR review

Anyone in the community is free to review the PR once the tests have passed.
Before you start reviewing make sure you have read Review guidelines. In short, see the following bullet-list:

Is this pull request ready for review? (if not, please submit in draft mode)
Check that all items from Before submitting are resolved
Make sure the title is self-explanatory and the description concisely explains the PR
Add labels and milestones (and optionally projects) to the PR so it can be classified; Bugfixes should be including in bug-fix release milestones (m.f.X) and features should be included in (m.X.b) releases.

Did you have fun?

Make sure you had fun coding 🙃

pep8speaks · 2020-11-25T22:26:58Z

Hello @rohitgr7! Thanks for updating this PR.

There are currently no PEP 8 issues detected in this Pull Request. Cheers! 🍻

Comment last updated at 2020-12-14 17:27:54 UTC

codecov · 2020-11-25T22:42:13Z

Codecov Report

Merging #4861 (dfd47dc) into master (fde972f) will decrease coverage by 0%.
The diff coverage is 100%.

@@          Coverage Diff           @@
##           master   #4861   +/-   ##
======================================
- Coverage      93%     93%   -0%     
======================================
  Files         134     134           
  Lines        9907    9905    -2     
======================================
- Hits         9206    9204    -2     
  Misses        701     701

rohitgr7 · 2020-11-26T19:49:17Z

Thanks @carmocca for the review :)

carmocca · 2020-11-26T19:51:30Z

Unrelated to the PR but, is there a reason for version starting at 0?

file.ckpt  # I find this confusing, kind of implies this is -v-1
file-v0.ckpt
file-v1.ckpt
...

instead of:

file.ckpt
file-v1.ckpt
file-v2.ckpt
...

Borda · 2020-11-26T22:14:46Z

Unrelated to the PR but, is there a reason for version starting at 0?

there is no special reason :]
cc: @PyTorchLightning/core-contributors

rohitgr7 · 2020-12-02T19:53:42Z

@carmocca wanna give this a shot? #4861 (comment)

carmocca · 2020-12-02T19:56:17Z

@carmocca wanna give this a shot? #4861 (comment)

Sure, I will. Most likely this weekend

* disable version if not required * disable version if not required * pep * chlog * improve test * improve test * parametrize test and update del_list * Update pytorch_lightning/callbacks/model_checkpoint.py Co-authored-by: Carlos Mocholí <carlossmocholi@gmail.com> * try appending version to already saved ckpt_file * Revert "try appending version to already saved ckpt_file" This reverts commit 710e05e. * add more assertions * use BoringModel Co-authored-by: Carlos Mocholí <carlossmocholi@gmail.com> Co-authored-by: chaton <thomas@grid.ai> Co-authored-by: Roger Shieh <sh.rog@protonmail.ch>

rohitgr7 changed the title ~~Bugfix/ckpt fname~~ Bugfix/ckpt fname [ci skip] Nov 25, 2020

rohitgr7 force-pushed the bugfix/ckpt_fname branch from 11952d6 to 9464b16 Compare November 26, 2020 18:04

rohitgr7 changed the title ~~Bugfix/ckpt fname [ci skip]~~ Fix saved filename in ModelCheckpoint if it's already present Nov 26, 2020

rohitgr7 changed the title ~~Fix saved filename in ModelCheckpoint if it's already present~~ Fix saved filename in ModelCheckpoint if it already exists Nov 26, 2020

rohitgr7 added bug Something isn't working checkpointing Related to checkpointing labels Nov 26, 2020

rohitgr7 added this to the 1.0.x milestone Nov 26, 2020

rohitgr7 marked this pull request as ready for review November 26, 2020 18:32

rohitgr7 requested review from Borda, SeanNaren, ananyahjha93, awaelchli, justusschock, nateraw, tchaton, teddykoker and williamFalcon as code owners November 26, 2020 18:32