GitLab

* speed up the AUC metric by parallel sort.

* fix unix compile.

* fix template in unix.

* remove warning.

warning: type qualifiers ignored on function return type
[-Wignored-qualifiers]

* auto replace gbdt::prediction

* add test if_else_prediction

* not override gbdt_prediction.cpp

* close ifstream

* re-order .travis.yml

* translate model to if-else

* support multiclass and predictleaf

* remove java option for now

* support multi-thread

* add task:convert_model

* fixed a bug in BinMapper: division by zero

* Revert "fixed a bug in BinMapper: division by zero"

This reverts commit 228380284f6a9ba174982e210e14d0aecde8e837.

* added CHECK for min_data_in_bin and fixed typo

On some OpenCL platforms the GPU code may fail to compile.
This patch fixes the previously non-working exception handler
under this failure case and prints a nice build log.

* Fix warnings when compiled with -pedantic

* add -DBOOST_ALL_NO_LIB for windows build

* fix some more MSVC warnings

* Break OpenCL string literal to smaller pieces to avoid error C2026 of MSVC

The string was longer than the limit of 16380 single-byte characters.

This affects Visual Studio 2005 - 2015. Untested on VS 2017.

* change whitespace to underline in feature names

* add test

* fix bug

* fix bug

* warning -> fatal

* Revert "python-package: support valid_names in scikit-learn API (#420)"

This reverts commit de39dbcf.

* Revert "faster histogram sum up (#418)"

This reverts commit 98c7c2a3.

* some refactor.

* two stage sum up to reduce sum up error.

* add more two-stage sumup.

* some refactor.

* add alignment.

* change name to aligned_allocator.

* remove some useless sumup.

* fix a warning.

* add -march=native .

* remove the padding of gradients.

* no alignment.

* fix test.

* change KNumSumupGroup to 32768.

* change gcc flags.

* RMSE (L2) -> MSE (true L2)

* Remove sqrt unneeded reference

* Square L2 test (RMSE to MSE)

* No square root on test

* Attempt to add RMSE

* refine prediction logic.

* fix test.

* fix out_len in training score of Dart.

* improve predict speed for high dimension data.

* try use unordered_map for sparse prediction.

* avoid using unordered_map.

* clean code.

* fix test.

* move predict buffer to Predictor.

* refine prediction logic.

* fix test.

* fix out_len in training score of Dart.

* improve predict speed for high dimension data.