v25.10.0 release
Release notes as follows:
- Updates RAPIDS dependencies to 25.10
- Aligns LinearRegression standardization behavior with Spark MLlib
- Adds experimental support for system allocated memory (SAM) with headroom for improved out-of-core performance on systems with high speed C2C link between CPU and GPU (e.g. GH200).
- Adds support for cuda 13.
Known issues:
- CrossValidator for RandomForest over Spark Connect will fail in Spark 4.0. Fix pending in Spark 4.1
- NearestNeighbors and DBSCAN are broken in this release, with a fix pending in cuML and RAFT 25.12. Please use the previous 25.08 release for these algorithms until 25.12.
pip package available at https://pypi.org/project/spark-rapids-ml/25.10.0/