Граф коммитов

417 Коммитов

Автор SHA1 Сообщение Дата
Carlos Becker 9595b8c11a Fixed library install dir on Windows 2017-01-24 09:58:08 +01:00
Yachen Yan bc628ee4b8 Parse a LightGBM model json dump (#253)
* Add lgb.Booster parsing function

* Add data.table, magrittr, jsonlite to Imports
2017-01-24 16:10:35 +08:00
cbecker 38ea6f616a Fixed compilation without OpenMP (#255) 2017-01-24 16:10:25 +08:00
Guolin Ke 6a367ac41c fix bug for subset 2017-01-24 13:44:43 +08:00
Guolin Ke 3489607fbc reduce function call cost for constructing subset. 2017-01-24 11:44:24 +08:00
Guolin Ke 6c736da932 fix error that sum_hessian may be zero. 2017-01-23 20:16:27 +08:00
wxchan a4a0235d17 use json instead of repr/eval for pandas_categorical (#247)
* use json instead of repr/eval for pandas_categorical

* fix json dumps with numpy data

* add more test cases
2017-01-23 19:06:36 +08:00
Guolin Ke 9c5dbdde5c [bug fix] fix predict sigmoid; fix bagging bug. 2017-01-23 19:04:38 +08:00
Guolin Ke fd0cbe6587 slight improve prediction speed 2017-01-23 16:30:20 +08:00
Guolin Ke 4948dcc5f1 improve bagging speed 2017-01-23 15:09:25 +08:00
Guolin Ke c8fbd42b61 use subset to speed up bagging 2017-01-22 20:54:02 +08:00
Guolin Ke 873528c1f8 fix complie error in vs2013 2017-01-22 15:42:35 +08:00
Yachen Yan 8da60f68b8 fix the reshape and predleaf arguments (#244) 2017-01-22 10:34:03 +08:00
Guolin Ke 112ab6f9b6 fix one bug in NextInt 2017-01-21 17:18:11 +08:00
wxchan aeeef276a1 [WIP] expose 'load/save model from/to string' to c_api (#241)
* expose save model to string to c_api

* add 'save-model-to-string' to python; use it to copy model

* remove boosting_type
2017-01-21 14:08:49 +08:00
cbecker 7dca0bb2a4 Add `feature_infos` to LoadModelFromString() (#239) 2017-01-20 16:39:22 +08:00
Guolin Ke 7426ac3cbd add map metric 2017-01-20 15:30:49 +08:00
wxchan e9b82412ee fix .gitignore (#238) 2017-01-20 14:32:30 +08:00
wxchan abaefb54ce [python-package] add plot importance (#237)
* add plot importance

* add plot example
2017-01-20 13:53:18 +08:00
Guolin Ke 46d4eecf2e change default sigmoid to 1.0f, slight improve binary accuracy 2017-01-20 11:25:20 +08:00
Tsukasa OMOTO d57118ba02 Fix warning (#236) 2017-01-19 21:59:08 +08:00
cbecker c0f8e38d66 Added `SaveModelToString` and modified `SaveModelToFile` to use it. (#235) 2017-01-19 18:41:32 +08:00
cbecker df358b2dbb Added success return value to LoadFileToBoosting and SaveModelToFile (#234) 2017-01-19 18:24:36 +08:00
wxchan 82fcfa0e57 fix warning; fix doc bug (#232) 2017-01-17 16:40:03 +08:00
Laurae a2ae8384d1 More elegant, more obvious error for the enduser (#229) 2017-01-17 10:25:30 +08:00
Laurae 54a2356c62 Attempt to fix categorical issue in R package (#228)
* Attempt to fix categorical issue in R package

* Add small fix for categorical IDs

* Add error for unknown numbers/columns
2017-01-17 00:05:40 +08:00
wxchan 57d552726f fix bug for pandas auto categorical_feature (#218)
* fix bug for categorical_feature

* add test on load model with categorical feature

* add unseen category in test dataset

* save/load pandas_categorical to model

* fix logic

* cast pandas columns to string

* add load pandas_categorical from file to _InnerPredictor init
2017-01-16 17:01:48 +08:00
wxchan adb8fb4e5c refine callback (#223) 2017-01-16 16:34:22 +08:00
wxchan 8c2341b31c fix sklearn interface compat (#222) 2017-01-16 16:34:08 +08:00
Guolin Ke 3db4216a0b [R-package] update installation for OSX 2017-01-15 20:22:32 +08:00
Tyler Hunt e7431447e1 added a little documentation (#210) 2017-01-14 23:32:31 +08:00
wxchan 2e100eeb0d not check metrics w/o eval set (#212) 2017-01-14 19:27:48 +08:00
Guolin Ke 011fe02487 output more feature information in model file, #192 2017-01-14 12:56:52 +08:00
Guolin Ke 9fd5bc2545 fix bug for get feature name from freed dataset 2017-01-14 12:30:33 +08:00
Guolin Ke 7dec4dec6a [R-package] update document for OSX installation 2017-01-13 13:02:46 +08:00
Guolin Ke 4ee993c4db Update README.md 2017-01-13 12:57:40 +08:00
Guolin Ke 929443316c [R-package] fix DllExport 2017-01-13 11:17:11 +08:00
Guolin Ke d72d935929 change warning message for wrong parameter type. 2017-01-13 10:53:34 +08:00
cbecker 300cee50b0 Make openmp optional, allow for dylib mac library output (#204)
* Added possibility to compile without openmp (useful for clang on Mac)

* Added CMake option to create dylib on MacOS
2017-01-12 22:46:09 +08:00
cbecker 46818ac60f Export low level interface API for MSVC (#203) 2017-01-12 19:59:43 +08:00
Qiwei Ye e5b94fbe7d Merge pull request #200 from ClimbsRocks/master
sklearn compatibility update- renames .feature_importance_ to .featur…
2017-01-12 17:16:20 +08:00
Qiwei Ye 859f6ed08a Merge pull request #201 from ClimbsRocks/patch-2
minor typo fixes
2017-01-12 17:03:57 +08:00
Preston Parry 3c9cdd18d8 minor typo fixes 2017-01-12 00:35:43 -08:00
ClimbsRocks b0f7aa508a sklearn compatibility update- renames .feature_importance_ to .feature_importances_ 2017-01-12 00:01:37 -08:00
wxchan 6c248d37c3 suppprt pandas categorical (#193)
* suppprt pandas categorical

* refine logic

* make default=auto

* fix train/valid categorical codes

* add test

* unify set _predictor

* fix tests

* fix warning

* support feature_name=int
2017-01-12 15:27:35 +08:00
Preston Parry 00e5b2447a Minor fixes and clarifications (#197) 2017-01-12 15:27:11 +08:00
Laurae cadcdf615d Fix R package compilation error (#194) 2017-01-12 01:03:07 +08:00
Guolin Ke 455ba60fb5 [R-package] separate the verbose and record 2017-01-11 11:06:02 +08:00
Guolin Ke 0ddb3efd49 [R-package] change ```startsWith``` to ```grepl```. 2017-01-11 08:26:20 +08:00
Kirill Sevastyanenko 2035c54b3b Stylistic changes r package (#184)
* src & callbacks

* lgb.Booster and utils

* cv

* wip lgb.Dataset

* lgb.Dataset

* lgb.Predictor

* lgb.train

* typos

* add flags to facilitate macosx compilation

* fix basic_string template error with clang

* most unfortunate mode of development

* fixup tests

* last test

* roxygen

* roxygen v5.x.x
2017-01-11 08:20:00 +08:00