Helper for handling categorical codes.

Initialize DMatrix from cupy ndarray.

Generate predict parameters which will be passed into xgboost.XGBModel().predict()

Get the gpu id from the task resources

Gets the current max number of concurrent tasks.

Get rabit context arguments to send to each worker.

Get the xgboost.sklearn.XGBModel default parameters and filter out some

Return dataframe for prediction when applicable.

Decorate a metric function from sklearn. Converts an metric function that uses the typical sklearn metric signature so that it is compatible

Return the bool to indicate if it's a single prediction, true is single prediction, and the returned type of the user-defined function. The va

Prepare the input including column pruning, repartition and so on

Print summary of certain result map.

Reference data from CUDA array interface.

Test constructing DMatrix from cudf

Test constructing DMatrix from cupy

It handles 1. Convert vector type to array type 2. Cast to Array(Float32)

Values in feature columns must be integral types or float/double types

Function called after training iteration.

Function called after training.

Redirect to booster attr.

Function called before training iteration.

Function called before training.

NOLINT

Subprocess run wrapper.

Check string inputs.

Check the quantile cut getter.

Check with optional cupy.

Validate there's no inf in X.

Test init estimation.

Check error message for inconsistent feature types.

Test for precision with ranking and classification.

Test for the `quantile` loss.

Test for quantile loss.

Check obtaining the same cut values given a reference.

Issue #9271, not every worker has missing value.

Tests for having irregular data shapes.

Print the message via the communicator. @param msg @throws XGBoostError