[tfjs-layers[ Fix Layer mem leak for 1D targets #5988
Merged
+41
−2
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Fixed #5979
The memory leak happens when the target input for fit method is a 1D tensor, and the train requires evaluation split.
A reshaped the tensor will be created for the targets if the target tensor is 1D tensor, but the training will miss disposing it when eval split is required.
Also fixed the test describe predicates when flags does not existing on testEnvs object.
To see the logs from the Cloud Build CI, please join either our discussion or announcement mailing list.
This change is