-
Notifications
You must be signed in to change notification settings - Fork 3.4k
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.
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
Reset the current progress tracking state during double evaluation #11119
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
carmocca
added
fault tolerance
progress tracking (internal)
Related to the progress tracking dataclasses
experimental
trainer: validate
trainer: test
trainer: predict
labels
Dec 17, 2021
carmocca
commented
Dec 17, 2021
carmocca
force-pushed
the
bugfix/back-to-back-eval
branch
from
December 17, 2021 16:05
1e77718
to
942f0ec
Compare
carmocca
requested review from
awaelchli,
Borda,
justusschock,
kaushikb11,
rohitgr7,
SeanNaren,
tchaton and
williamFalcon
as code owners
December 17, 2021 16:15
Build Error! No Linked Issue found. Please link an issue or mention it in the body using #<issue_id> |
justusschock
approved these changes
Dec 17, 2021
tchaton
approved these changes
Dec 17, 2021
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM !
Borda
approved these changes
Dec 17, 2021
awaelchli
added a commit
that referenced
this pull request
Dec 21, 2021
…11119) Co-authored-by: Carlos Mocholi <[email protected]>
awaelchli
added a commit
that referenced
this pull request
Dec 21, 2021
…11119) Co-authored-by: Carlos Mocholi <[email protected]>
11 tasks
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Labels
experimental
fault tolerance
progress tracking (internal)
Related to the progress tracking dataclasses
ready
PRs ready to be merged
trainer: predict
trainer: test
trainer: validate
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What does this PR do?
With fault-tolerance disabled, the following runs twice through the test set, it's equivalent to
max_test_epochs=2
However, currently in master, if fault-tolerance is enabled, the second test exits early. This is because the
done
condition for the test loop has been already fulfilled:To avoid this, we need to check whether we are NOT fitting, and whether the previous state is done, then in that case reset the state.
This is different from how fit works where this is the expected behavior and the user is required to set a different
max_epochs
value. The equivalent would be that the user has to changeself.num_dataloaders
which does not make sense.Predict also has this bug, but there we don't need the fit check.
Test plan
This is covered by existing tests, for example:
test_checkpoint_path_input
Does your PR introduce any breaking changes? If yes, please list them.
None
Before submitting
PR review
cc @carmocca @justusschock @awaelchli @ninginthecloud @Borda