243 Commits

Author SHA1 Message Date
amdsc21
c50cc424bc sync Mar 27 2023 2023-03-27 18:54:41 +02:00
Jiaming Yuan
acc110c251
[MT-TREE] Support prediction cache and model slicing. (#8968)
- Fix prediction range.
- Support prediction cache in mt-hist.
- Support model slicing.
- Make the booster a Python iterable by defining `__iter__`.
- Cleanup removed/deprecated parameters.
- A new field in the output model `iteration_indptr` for pointing to the ranges of trees for each iteration.
2023-03-27 23:10:54 +08:00
amdsc21
7ee4734d3a rm device_helpers.hip.h from cu 2023-03-26 00:24:11 +01:00
amdsc21
1474789787 add new file 2023-03-25 04:54:02 +01:00
amdsc21
7fbc561e17 initial merge 2023-03-25 04:31:55 +01:00
Jiaming Yuan
5891f752c8
Rework the MAP metric. (#8931)
- The new implementation is more strict as only binary labels are accepted. The previous implementation converts values greater than 1 to 1.
- Deterministic GPU. (no atomic add).
- Fix top-k handling.
- Precise definition of MAP. (There are other variants on how to handle top-k).
- Refactor GPU ranking tests.
2023-03-22 17:45:20 +08:00
Jiaming Yuan
a093770f36
Partitioner for multi-target tree. (#8922) 2023-03-16 18:49:34 +08:00
Jiaming Yuan
26209a42a5
Define git attributes for renormalization. (#8921) 2023-03-16 02:43:11 +08:00
amdsc21
8207015e48 fix ../tests/cpp/common/test_span.h 2023-03-14 22:19:06 +01:00
Jiaming Yuan
8be6095ece
Implement NDCG cache. (#8893) 2023-03-13 22:16:31 +08:00
amdsc21
b4dbe7a649 fix isnan 2023-03-11 02:39:58 +01:00
amdsc21
3a07b1edf8 complete test porting 2023-03-11 02:17:05 +01:00
amdsc21
f0b8c02f15 merge latest changes 2023-03-10 22:10:20 +01:00
Jiaming Yuan
46dfcc7d22
Define a new ranking parameter. (#8887) 2023-03-09 17:46:24 +08:00
amdsc21
ed45aa2816 Merge branch 'master' into dev-hui 2023-03-08 00:39:33 +01:00
Jiaming Yuan
f236640427
Support F order for the tensor type. (#8872)
- Add F order support for tensor and view.
- Use parameter pack for automatic type cast. (avoid excessive static cast for shape).
2023-03-08 03:27:49 +08:00
amdsc21
c51a1c9aae rename hip.cc to hip 2023-03-07 05:39:53 +01:00
amdsc21
6039a71e6c add hip structure 2023-03-07 02:17:19 +01:00
Jiaming Yuan
4d665b3fb0
Restore clang tidy test. (#8861) 2023-03-03 13:47:04 -08:00
Jiaming Yuan
cce4af4acf
Initial support for quantile loss. (#8750)
- Add support for Python.
- Add objective.
2023-02-16 02:30:18 +08:00
Jiaming Yuan
282b1729da
Specify the number of threads for parallel sort. (#8735)
* Specify the number of threads for parallel sort.

- Pass context object into argsort.
- Replace macros with inline functions.
2023-02-16 00:20:19 +08:00
Jiaming Yuan
31d3ec07af
Extract device algorithms. (#8789) 2023-02-13 20:53:53 +08:00
Jiaming Yuan
457f704e3d
Add quantile metric. (#8761) 2023-02-13 19:07:40 +08:00
Rong Ou
ed91e775ec
Fix quantile tests running on multi-gpus (#8775)
* Fix quantile tests running on multi-gpus

* Run some gtests with multiple GPUs

* fix mgpu test naming

* Instruct NCCL to print extra logs

* Allocate extra space in /dev/shm to enable NCCL

* use gtest_skip to skip mgpu tests

---------

Co-authored-by: Hyunsu Philip Cho <chohyu01@cs.washington.edu>
2023-02-12 17:00:26 -08:00
Jiaming Yuan
17b709acb9
Rename ranking utils to threading utils. (#8785) 2023-02-12 05:41:18 +08:00
Jiaming Yuan
5f76edd296
Extract make metric name from ranking metric. (#8768)
- Extract the metric parsing routine from ranking.
- Add a test.
- Accept null for string view.
2023-02-09 18:30:21 +08:00
Jiaming Yuan
48cefa012e
Support multiple alphas for segmented quantile. (#8758) 2023-02-07 17:17:59 +08:00
Jiaming Yuan
28bb01aa22
Extract optional weight. (#8747)
- Extract optional weight from coommon.h to reduce dependency on this header.
- Add test.
2023-02-07 03:11:53 +08:00
Rong Ou
66191e9926
Support cpu quantile sketch with column-wise data split (#8742) 2023-02-05 14:26:24 +08:00
Jiaming Yuan
c1786849e3
Use array interface for CSC matrix. (#8672)
* Use array interface for CSC matrix.

Use array interface for CSC matrix and align the interface with CSR and dense.

- Fix nthread issue in the R package DMatrix.
- Unify the behavior of handling `missing` with other inputs.
- Unify the behavior of handling `missing` around R, Python, Java, and Scala DMatrix.
- Expose `num_non_missing` to the JVM interface.
- Deprecate old CSR and CSC constructors.
2023-02-05 01:59:46 +08:00
Jiaming Yuan
3760cede0f
Consistent use of context to specify number of threads. (#8733)
- Use context in all tests.
- Use context in R.
- Use context in C API DMatrix initialization. (0 threads is used as dft).
2023-01-30 15:25:31 +08:00
Rong Ou
8af98e30fc
Use in-memory communicator to test quantile (#8710) 2023-01-27 23:28:28 +08:00
Jiaming Yuan
4416452f94
Return single thread from context when called inside omp region. (#8693) 2023-01-18 09:23:37 +08:00
Jiaming Yuan
43152657d4
Extract JSON type check. (#8677)
- Reuse it in `GetMissing`.
- Add test.
2023-01-17 03:11:07 +08:00
Jiaming Yuan
cfa994d57f
Multi-target support for L1 error. (#8652)
- Add matrix support to the median function.
- Iterate through each target for quantile computation.
2023-01-11 05:51:14 +08:00
Jiaming Yuan
8d545ab2a2
Implement fit stump. (#8607) 2023-01-04 04:14:51 +08:00
Rong Ou
3ceeb8c61c
Add data split mode to DMatrix MetaInfo (#8568) 2022-12-25 20:37:37 +08:00
Jiaming Yuan
38887a1876
Fix windows build on buildkite. (#8602) 2022-12-16 21:12:24 +08:00
Jiaming Yuan
43a647a4dd
Fix inference with categorical feature. (#8591) 2022-12-15 17:57:26 +08:00
Jiaming Yuan
3e26107a9c
Rename and extract Context. (#8528)
* Rename `GenericParameter` to `Context`.
* Rename header file to reflect the change.
* Rename all references.
2022-12-07 04:58:54 +08:00
Jiaming Yuan
e3bf5565ab
Extract transform iterator. (#8498) 2022-12-05 21:37:07 +08:00
Rong Ou
8e76f5f595
Use DataSplitMode to configure data loading (#8434)
* Use `DataSplitMode` to configure data loading
2022-11-08 16:21:50 +08:00
Jiaming Yuan
3ef1703553
Allow using string view to find JSON value. (#8332)
- Allow comparison between string and string view.
- Fix compiler warnings.
2022-10-13 17:10:13 +08:00
Rong Ou
668b8a0ea4
[Breaking] Switch from rabit to the collective communicator (#8257)
* Switch from rabit to the collective communicator

* fix size_t specialization

* really fix size_t

* try again

* add include

* more include

* fix lint errors

* remove rabit includes

* fix pylint error

* return dict from communicator context

* fix communicator shutdown

* fix dask test

* reset communicator mocklist

* fix distributed tests

* do not save device communicator

* fix jvm gpu tests

* add python test for federated communicator

* Update gputreeshap submodule

Co-authored-by: Hyunsu Philip Cho <chohyu01@cs.washington.edu>
2022-10-05 14:39:01 -08:00
Rory Mitchell
d686bf52a6
Reduce time for some multi-gpu tests (#8288)
* Faster dask tests

* Reuse AllReducer objects in tests.

* Faster boost from prediction tests.

* Use rmm dask fixture.

* Speed up dask demo.

* mypy

* Format with black.

* mypy

* Clang-tidy

Co-authored-by: Hyunsu Philip Cho <chohyu01@cs.washington.edu>
2022-10-04 02:49:33 -08:00
Jiaming Yuan
6d1452074a
Remove MGPU cpp tests. (#8276)
Co-authored-by: Hyunsu Philip Cho <chohyu01@cs.washington.edu>
2022-09-27 21:18:23 +08:00
Rory Mitchell
8f77677193
Use quantised gradients in gpu_hist histograms (#8246) 2022-09-26 17:35:35 +02:00
Jiaming Yuan
fffb1fca52
Calculate base_score based on input labels for mae. (#8107)
Fit an intercept as base score for abs loss.
2022-09-20 20:53:54 +08:00
Jiaming Yuan
bc818316f2
Prepare for improving Windows networking compatibility. (#8234)
* Prepare for improving Windows networking compatibility.

* Include dmlc filesystem indirectly as dmlc/filesystem.h includes windows.h, which
  conflicts with winsock2.h
* Define `NOMINMAX` conditionally.
* Link the winsock library when mysys32 is used.
* Add config file for read the doc.
2022-09-10 15:16:49 +08:00
Jiaming Yuan
441ffc017a
Copy data from Ellpack to GHist. (#8215) 2022-09-06 23:05:49 +08:00