zhengruifeng opened a new pull request, #57715: URL: https://github.com/apache/spark/pull/57715
### What changes were proposed in this pull request? This follow-up to SPARK-58250 excludes `OneVsRestModel.classifier` from parameter-metadata sizing. It continues to account explicitly for label metadata and the fitted child models. The regression test initializes the source classifier logger before fitting OneVsRest, ensuring the estimate does not retain shared runtime state through the classifier param. ### Why are the changes needed? The fitted OneVsRestModel retains the base classifier estimator in its param map. That estimator can retain shared runtime state, including SparkSession-related state, which must not contribute to Spark Connect ML model-cache accounting. ### Does this PR introduce _any_ user-facing change? No. ### How was this patch tested? Updated the OneVsRestModel estimated-size regression test. Static checks: `git diff --check`, non-ASCII scan, and changed-Scala line-length scan. The targeted Scala suite was not run locally. ### Was this patch authored or co-authored using generative AI tooling? Generated-by: Codex GPT-5 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
