Dmitry Razdoburdin
2179baa50c
[SYC]. Implementation of HostDeviceVector ( #10842 )
2024-09-25 04:45:17 +08:00
Jiaming Yuan
bc69a3e877
[EM] Improve memory estimation for quantile sketching. ( #10843 )
...
I- Add basic estimation for RMM.
- Re-estimate after every sub-batch.
- Some debug logs for memory usage.
- Fix the locking mechanism in the memory allocator logger.
2024-09-25 03:20:09 +08:00
Bobby Wang
f3df0d0eb4
[jvm-packages] update scala style configuration ( #10836 )
2024-09-24 17:39:44 +08:00
Bobby Wang
2a03685bff
[jvm-packages] shade xgboost spark packages ( #10833 )
2024-09-24 15:46:06 +08:00
Jiaming Yuan
68a8865bc5
[CI] Fix PyLint errors. ( #10837 )
2024-09-24 14:09:32 +08:00
Bobby Wang
982ee34658
[jvm-packages] fix surefire ( #10835 )
2024-09-24 14:08:30 +08:00
Jiaming Yuan
e228c1a121
[EM] Make page concatenation optional. ( #10826 )
...
This PR introduces a new parameter `extmem_concat_pages` to make the page concatenation optional for GPU hist. In addition, the document is updated for the new GPU-based external memory.
2024-09-24 06:19:28 +08:00
jonibr22
215da76263
[R] Fix xgb.model.dt.tree in case where all leaves are negative ( #10798 )
2024-09-24 05:02:33 +08:00
Bobby Wang
19b55b300b
[jvm-packages] Support Ranker ( #10823 )
2024-09-22 02:02:15 +08:00
Dmitry Razdoburdin
d7599e095b
[SYCL] Add dask support for distributed ( #10812 )
2024-09-22 02:01:57 +08:00
Jiaming Yuan
2a37a8880c
Check correct dump format for gblinear. ( #10831 )
2024-09-21 00:32:52 +08:00
Jiaming Yuan
24241ed6e3
[EM] Compress dense ellpack. ( #10821 )
...
This helps reduce the memory copying needed for dense data. In addition, it helps reduce memory usage even if external memory is not used.
- Decouple the number of symbols needed in the compressor with the number of features when the data is dense.
- Remove the fetch call in the `at_end_` iteration.
- Reduce synchronization and kernel launches by using the `uvector` and ctx.
2024-09-20 18:20:56 +08:00
Jiaming Yuan
d5e1c41b69
[coll] Use loky for rabit op tests. ( #10828 )
2024-09-20 16:46:05 +08:00
Valentin Waeselynck
15c6172e09
[doc] Improve the model introduction. ( #10822 )
2024-09-19 02:33:49 +08:00
Jiaming Yuan
96bbf80457
[EM] Suport quantile objectives for GPU-based external memory. ( #10820 )
...
- Improved error message for memory usage.
- Support quantile-based objectives for GPU external memory.
2024-09-17 13:27:02 +08:00
shlomota
de00e07087
Fix misleading error when feature names are missing during inference ( #10814 )
2024-09-13 23:30:50 +08:00
Bobby Wang
67c8c96784
[jvm-packages] [breaking] rework xgboost4j-spark and xgboost4j-spark-gpu ( #10639 )
...
- Introduce an abstract XGBoost Estimator
- Update to the latest XGBoost parameters
- Add all XGBoost parameters supported in XGBoost4j-spark.
- Add setter and getter for these parameters.
- Remove the deprecated parameters
- Address the missing value handling
- Remove any ETL operations in XGBoost
- Rework the GPU plugin
- Expand sanity tests for CPU and GPU consistency
2024-09-11 15:54:19 +08:00
Jiaming Yuan
d94f6679fc
[EM] Avoid synchronous calls and unnecessary ATS access. ( #10811 )
...
- Pass context into various functions.
- Factor out some CUDA algorithms.
- Use ATS only for update position.
2024-09-10 14:33:14 +08:00
Jiaming Yuan
ed5f33df16
[EM] Multi-level quantile sketching for GPU. ( #10813 )
2024-09-10 13:08:34 +08:00
Jiaming Yuan
3ef8383d93
[doc] Fix custom_metric_obj.rst [skip ci] ( #10796 ) ( #10815 )
...
Added the square to the derivative in the hessian
Co-authored-by: Corentin Santos <corentin.santos@iphc.cnrs.fr>
2024-09-10 05:11:43 +08:00
Dmitry Razdoburdin
bba6aa74fb
[SYCL] Fix for sycl support with sklearn estimators ( #10806 )
...
---------
Co-authored-by: Dmitry Razdoburdin <>
2024-09-09 14:14:07 +08:00
Jiaming Yuan
5f7f31d464
[EM] Refactor ellpack construction. ( #10810 )
...
- Remove the calculation of n_symbols in the accessor.
- Pack initialization steps into the parameter list.
- Pass the context into various ctors.
- Specialization for dense data to prepare for further compression.
2024-09-09 14:10:10 +08:00
dependabot[bot]
c69c4adb58
Bump actions/setup-python from 5.1.1 to 5.2.0 ( #10768 )
...
Bumps [actions/setup-python](https://github.com/actions/setup-python ) from 5.1.1 to 5.2.0.
- [Release notes](https://github.com/actions/setup-python/releases )
- [Commits](39cd14951b...f677139bbe )
---
updated-dependencies:
- dependency-name: actions/setup-python
dependency-type: direct:production
update-type: version-update:semver-minor
...
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2024-09-08 00:09:22 +08:00
david-cortes
f52f11e1d7
[R] Allow passing data.frame to SHAP ( #10744 )
2024-09-02 19:44:12 +08:00
dependabot[bot]
ec8cfb3267
Bump actions/upload-artifact from 4.3.4 to 4.4.0 ( #10770 )
...
Bumps [actions/upload-artifact](https://github.com/actions/upload-artifact ) from 4.3.4 to 4.4.0.
- [Release notes](https://github.com/actions/upload-artifact/releases )
- [Commits](0b2256b8c0...50769540e7 )
---
updated-dependencies:
- dependency-name: actions/upload-artifact
dependency-type: direct:production
update-type: version-update:semver-minor
...
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2024-09-02 17:52:32 +08:00
david-cortes
15b72571f3
[R] update serialization advise for new xgboost class ( #10794 )
2024-09-02 02:46:11 +08:00
dependabot[bot]
4f88ada219
Bump actions/setup-java from 4.2.1 to 4.2.2 ( #10769 )
...
Bumps [actions/setup-java](https://github.com/actions/setup-java ) from 4.2.1 to 4.2.2.
- [Release notes](https://github.com/actions/setup-java/releases )
- [Commits](99b8673ff6...6a0805fcef )
---
updated-dependencies:
- dependency-name: actions/setup-java
dependency-type: direct:production
update-type: version-update:semver-patch
...
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2024-09-02 02:35:48 +08:00
Samuel Marks
4503555274
POSIX compliant poll.h and mmap over sys/poll.h and mmap64 ( #10767 )
2024-09-01 15:47:30 +08:00
Jiaming Yuan
e1a2c1bbb3
[EM] Merge GPU partitioning with histogram building. ( #10766 )
...
- Stop concatenating pages if there's no subsampling.
- Use a single iteration for histogram build and partitioning.
2024-08-31 03:25:37 +08:00
Jiaming Yuan
98ac153265
Avoid warning from NVCC. ( #10757 )
2024-08-30 16:11:31 +08:00
Jiaming Yuan
5cc7c735e5
Don't link gputreeshap. ( #10758 )
...
Co-authored-by: Philip Hyunsu Cho <chohyu01@cs.washington.edu>
2024-08-30 14:40:58 +08:00
Jiaming Yuan
34d4ab455e
[EM] Avoid stream sync in quantile sketching. ( #10765 )
...
.
2024-08-30 12:33:24 +08:00
Jiaming Yuan
61dd854a52
[EM] Refactor GPU histogram builder. ( #10764 )
...
- Expose the maximum number of cached nodes to be consistent with the CPU implementation. Also easier for testing.
- Extract the subtraction trick for easier testing.
- Split up the `GradientQuantiser` to avoid circular dependency.
2024-08-30 02:39:14 +08:00
Jiaming Yuan
34937fea41
[EM] Python wrapper for the ExtMemQuantileDMatrix. ( #10762 )
...
Not exposed to the document yet.
- Add C API.
- Add Python API.
- Basic CPU tests.
2024-08-29 04:08:25 +08:00
Jiaming Yuan
7510a87466
[EM] Reuse the quantile container. ( #10761 )
...
Use the push method to merge the quantiles instead of creating multiple containers. This
reduces the memory usage by consistent pruning.
2024-08-29 01:39:55 +08:00
Jiaming Yuan
4fe67f10b4
[EM] Have one partitioner for each batch. ( #10760 )
...
- Initialize one partitioner for each batch.
- Collect partition size during initialization.
- Support base ridx in the finalization.
2024-08-29 01:35:17 +08:00
david-cortes
3043827efc
[R] Update vignette "XGBoost presentation" ( #10749 )
...
---------
Co-authored-by: Jiaming Yuan <jm.yuan@outlook.com>
2024-08-28 16:22:54 +08:00
Philip Hyunsu Cho
7794d3da8a
Ensure that pip check does not fail due to bad platform tag ( #10755 )
...
* Remove custom tag generation
* Revert "Remove custom tag generation"
This reverts commit fe3cf0e8786c7dc05e1deced3a1c92cd79094735.
* Fetch an accurate platform tag from Pip 22+
* Fix formatting
* TOML allows trailing commas
* Update patch
* Add trailing comma
* Fix up patch
* Use `packaging`
Co-authored-by: jakirkham <jakirkham@gmail.com>
---------
Co-authored-by: jakirkham <jakirkham@gmail.com>
2024-08-27 18:11:08 -07:00
Jiaming Yuan
64afe9873b
Increase timeout in C++ tests from 1 to 5 seconds. ( #10756 )
...
To avoid CI failures on FreeBSD.
2024-08-28 02:27:14 +08:00
Jiaming Yuan
bde1265caf
[EM] Return a full DMatrix instead of a Ellpack from the GPU sampler. ( #10753 )
2024-08-28 01:05:11 +08:00
Jiaming Yuan
d6ebcfb032
[EM] Support CPU quantile objective for external memory. ( #10751 )
2024-08-27 04:16:57 +08:00
david-cortes
12c6b7ceea
[R] Remove demos ( #10750 )
2024-08-27 04:16:36 +08:00
Jiaming Yuan
06c4246ff1
[CI] Workaround mypy errors. ( #10754 )
2024-08-27 02:54:11 +08:00
Jiaming Yuan
25966e4ba8
[EM] Pass batch parameter into extmem format. ( #10736 )
...
- Allow customization for format reading.
- Customize the number of pre-fetch batches.
2024-08-27 02:37:50 +08:00
Michael Mayer
074cad2343
[R] Finalizes switch to markdown doc ( #10733 )
...
---------
Co-authored-by: Jiaming Yuan <jm.yuan@outlook.com>
2024-08-27 01:25:06 +08:00
david-cortes
479ae8081b
[R] Add class names to coefficients ( #10745 )
2024-08-25 04:41:58 +08:00
Jiaming Yuan
fd0138c91c
[coll] Improve column split tests with named threads. ( #10735 )
2024-08-24 12:43:47 +08:00
Jiaming Yuan
55aef8f546
[EM] Avoid resizing host cache. ( #10734 )
...
* [EM] Avoid resizing host cache.
- Add SAM allocator and resource.
- Use page-based cache instead of stream-based cache.
2024-08-23 06:34:01 +08:00
James Lamb
dbfafd8557
[doc] Install the conda GPU variant in environments without CUDA ( #10731 )
2024-08-22 19:48:15 +08:00
Philip Hyunsu Cho
cd83fe6033
[breaking][CI] Use CTK 12.4 ( #10697 )
2024-08-21 19:59:34 -07:00