[AIR] Ensure that driver DatasetContext is propagated to the Trainer actor.
#29192
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
DatasetContext is propagated to the Trainer actor.
amogkam
reviewed
Oct 8, 2022
56d354a to
8c0c89c
Compare
amogkam
approved these changes
Oct 13, 2022
WeichenXu123
pushed a commit
to WeichenXu123/ray
that referenced
this pull request
Dec 19, 2022
…er` actor. (ray-project#29192) DatasetContext allows setting Datasets config values like the target max block size. Normally these are propagated from the driver to all tasks in the Dataset. However, when using the AIR trainer, context changes made by the driver don't get propagated to the training dataset preprocessor. I think this is likely because the BatchMapper doesn't save the context, so when we pass the BatchMapper to the worker that actually creates the final Dataset, the driver's context values are lost. Signed-off-by: Weichen Xu <weichen.xu@databricks.com>
8 tasks
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Outdated suggestions cannot be applied.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.
DatasetContext allows setting Datasets config values like the target max block size. Normally these are propagated from the driver to all tasks in the Dataset.
However, when using the AIR trainer, context changes made by the driver don't get propagated to the training dataset preprocessor. I think this is likely because the BatchMapper doesn't save the context, so when we pass the BatchMapper to the worker that actually creates the final Dataset, the driver's context values are lost.
Related issue number
Closes #29160
Checks
git commit -s) in this PR.scripts/format.shto lint the changes in this PR.