Use Lightning DataModules #130
Conversation
ananyahjha93
changed the title
Codecov Report
@@ Coverage Diff @@
## master #130 +/- ##
==========================================
+ Coverage 92.27% 92.30% +0.02%
==========================================
Files 79 79
Lines 4038 4000 -38
==========================================
- Hits 3726 3692 -34
+ Misses 312 308 -4
Flags with carried forward coverage won't be shown. Click here to find out more.
Continue to review full report at Codecov.
|
|
Good to go. Had to add some unrelated unittests because I couldn't stand the coverage failing (Autoencoder components, LARS optimizer) Currently pinned at lightning v0.9.0rc3. I would normally be against it, but this PR has a lot of changes which will probably cause some nasty merge conflicts, so I'll leave it up to you two if you want to merge before stable release of lightning 0.9.0. |
| @@ -19,6 +19,7 @@ def __init__( | |||
| data_dir, | |||
| val_split=5000, | |||
| num_workers=16, | |||
| batch_size=32, | |||
| *args, | |||
There was a problem hiding this comment.
this is tricky as it tends to confuse user why some arguments were not used...
| @@ -122,15 +124,6 @@ def configure_optimizers(self): | |||
| def prepare_data(self): | |||
| self.datamodule.prepare_data() | |||
There was a problem hiding this comment.
why do we still need to call prepare data? shouldn’t that be automatically called for the user?
There was a problem hiding this comment.
I think @nateraw mentioned here that we need manual calls to prepare_data for now: Lightning-AI/pytorch-lightning#2742
There was a problem hiding this comment.
solved here: Lightning-AI/pytorch-lightning#2755
so we can remove those calls to prepare_data across the repo as soon as this is merged/next release is up on PyPi
Before submitting
What does this PR do?
Resolves #129
I suggest we don't merge until 0.9.0 has been released. This PR uses the prerelease version 0.9.0rc2.
PR review
Anyone in the community is free to review the PR once the tests have passed.
If we didn't discuss your PR in Github issues there's a high chance it will not be merged.
Did you have fun?
Yeah 💯