GitLab

* Try to store best scores in R package Microsoft/LightGBM#686

* Add binding for best_score (unlock environment)

* Fix double comma

* Add best_score to lgb.Booster

* Remove CMAKE_CXX_COMPILER and CMAKE_C_COMPILER for OSX

* update .travis/test.sh

* update again

missing parameter "self" in signature of Booster.get_leaf_output

https://github.com/Microsoft/LightGBM/issues/705#issuecomment-315976423

* Created objectives and metrics xentropy and xentropy1

* Some coment and code cleanup.

* Added Kullback-Leibler version of metric. Changed some warning messages.

* Fixed sign error in KL-divergence calc.

* Removed __PRETTY_FUNCTION__.

* Fixed better name for alternative xentropy parameterization.
Documented details on the objectives / metrics in code comments.

* Common code for label interval checks. Cleanups.

* Use common utility for various weight property checks.

* Update key events to 15th July 2017

* Remove extra lines

* Add pypi to key events

* Change VS -> MinGW for CRAN

* Documentation to switch to MinGW by default

* Force cmake to run twice

* Try again dual build for Rtools

* Switch to cmake for building twice

* Try with Visual Studio as default, MinGW as fallback

* Try to remove VS appropriately when failing

* Attempt to get rid of build folder first

* Switch to unlink from rm rf

* Change lib_folder correctly when VS fails

* Add README updates from Microsoft/LightGBM@1c8355c

* Update default compiler doc and add GPU online install doc

* Better GPU doc

* add draft of RF.

* fix score bugs.

* fix scores.

* fix tests.

* update document

* fix GetPredictAt

* add keep_training_booster

* use model string

* reset handle; free dataset

thanks for building in early_stopping! super useful.

fixed english mistakes