Jiaming Yuan 0d0abe1845

Support optimal partitioning for GPU hist. (#7652) ...

* Implement `MaxCategory` in quantile.
* Implement partition-based split for GPU evaluation.  Currently, it's based on the existing evaluation function.
* Extract an evaluator from GPU Hist to store the needed states.
* Added some CUDA stream/event utilities.
* Update document with references.
* Fixed a bug in approx evaluator where the number of data points is less than the number of categories.

2022-02-15 03:03:12 +08:00

..

base.h

[breaking] Use integer atomic for GPU histogram. (#7180)

2021-08-28 05:17:05 +08:00

c_api.h

[breaking] Change internal model serialization to UBJSON. (#7556)

2022-01-16 02:11:53 +08:00

data.h

Prepare external memory support for hist. (#7638)

2022-02-10 16:58:02 +08:00

feature_map.h

Support more input types for categorical data. (#7220)

2021-09-16 20:39:30 +08:00

gbm.h

Remove omp_get_max_threads (#7608)

2022-01-28 16:09:22 +08:00

generic_parameters.h

Remove omp_get_max_threads in CPU predictor. (#7519)

2022-01-04 22:12:15 +08:00

global_config.h

Add use_rmm flag to global configuration (#6656)

2021-03-09 14:53:05 -08:00

host_device_vector.h

Implement extend method for meta info. (#5800)

2020-06-20 03:32:03 +08:00

intrusive_ptr.h

Implement ubjson. (#7549)

2022-01-10 23:24:23 +08:00

json_io.h

Implement ubjson. (#7549)

2022-01-10 23:24:23 +08:00

json.h

Implement new save_raw in Python. (#7572)

2022-01-19 02:27:51 +08:00

learner.h

Remove omp_get_max_threads (#7608)

2022-01-28 16:09:22 +08:00

linalg.h

[breaking] Change internal model serialization to UBJSON. (#7556)

2022-01-16 02:11:53 +08:00

linear_updater.h

Remove omp_get_max_threads in gbm and linear. (#7537)

2022-01-05 03:28:52 +08:00

logging.h

Add global configuration (#6414)

2020-12-03 00:05:18 -08:00

metric.h

Use double precision in metric calculation. (#7364)

2021-11-02 12:00:32 +08:00

model.h

Fix spelling in documents (#6948)

2021-05-11 20:44:36 +08:00

objective.h

Remove omp_get_max_threads in objective. (#7589)

2022-01-24 04:35:49 +08:00

parameter.h

Update dmlc-core submodule and conform to new API (#6431)

2021-01-05 16:12:22 -08:00

predictor.h

Remove omp_get_max_threads in CPU predictor. (#7519)

2022-01-04 22:12:15 +08:00

span.h

Fix span reverse iterator. (#7387)

2021-11-02 13:35:59 +08:00

string_view.h

Extract string view. (#7416)

2021-11-12 18:22:30 +08:00

task.h

Support optimal partitioning for GPU hist. (#7652)

2022-02-15 03:03:12 +08:00

tree_model.h

[breaking] Change internal model serialization to UBJSON. (#7556)

2022-01-16 02:11:53 +08:00

tree_updater.h

Remove omp_get_max_threads in tree updaters. (#7590)

2022-01-26 19:55:47 +08:00

version_config.h

Bump version to 1.6.0 in master. (#7259)

2021-10-07 16:09:26 +08:00