Skip to content
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

[breaking] [jvm-packages] Remove rabit check point. #9599

Merged
merged 7 commits into from
Sep 26, 2023

Conversation

trivialfis
Copy link
Member

@trivialfis trivialfis commented Sep 20, 2023

  • Add numBoostedRound to jvm packages
  • Remove rabit checkpoint version.
  • Change the starting version of training continuation in JVM [breaking].
  • Redefine the checkpoint version policy in jvm package. [breaking]
  • Rename the Python check point callback parameter. [breaking]
  • Unifies the checkpoint policy between Python and JVM.

The last item is a bit more subtle, the change aligns JVM packages with Python packages. After this PR, training continuation counts iteration from 0 instead of from the previous starting iteration. The checkpointing scheme is changed to use boosting versions. I don't understand the previous versioning scheme, likely a legacy inherited from rabit.

- Add `numBoostedRound` to jvm packages
- Remove rabit checkpoint version. [breaking]
- Change the starting version of training continuation in JVM [breaking].

The last item is a bit more subtle, the change aligns JVM packages with Python
packages. After this PR, the second training phrase counts iteration from 0 instead of
from the previous starting iteration.
@trivialfis trivialfis marked this pull request as ready for review September 22, 2023 13:19
@trivialfis
Copy link
Member Author

@wbo4958 Could you please help review this when you are available?

}

public void setVersion(int version) {
this.version = version;
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

seems still need to remove the definition of version

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Not sure what would it do to binary serialization.

booster.saveRabitCheckpoint();
}
for (int iter = 0; iter < numRounds; iter++) {
booster.update(dtrain, iter);
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

the original logic is

        if (obj != null) {
          booster.update(dtrain, obj);
        } else {
          booster.update(dtrain, iter);
        }

please double check

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Fixed. Need to create a test for the objective in a different PR.

@wbo4958
Copy link
Contributor

wbo4958 commented Sep 26, 2023

GLTM.

@trivialfis trivialfis merged commit c75a3bc into dmlc:master Sep 26, 2023
27 checks passed
@trivialfis trivialfis deleted the rabit-remove-chpt branch September 26, 2023 10:06
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants