XGBoost - Data too large

Description

Due to a technical limitation of the Java API of XGBoost, only data structures that fit into a 32-bit Java array (float[]) can be passed from H2O to the C++ XGBoost backend. We are working on an improved API that allows passing multiple arrays to handle arbitrarily large datasets. In the meantime, please fall back to H2O GBM for big datasets, or use the XGBoost algorithm through its original interfaces.

Assignee

Arno Candel

Reporter

Arno Candel

Labels

None

Affected Spark version

None

AffectedContact

None

AffectedCustomers

None

AffectedPilots

None

AffectedOpenSource

None

Support Assessment

None

Customer Request Type

None

Support ticket URL

None

End date

None

Baseline start date

None

Baseline end date

None

Task progress

None

Task mode

None

Priority

Major
Configure