Code

Add transfer delete shap#1

Code

Open

yuanborong wants to merge 158 commits intolav-patel:masterlav-patel/xgboost:masterfrom yuanborong:add_transfer_delete_SHAPyuanborong/xgboost:add_transfer_delete_SHAPCopy head branch name to clipboard

+10,112-3,312

Commits

Commits on Dec 24, 2019

• Add base margin to sklearn interface. (dmlc#5151)

  trivialfis

  authored
  0202e04

  Copy full SHA for 0202e04

• Update demo for ranking. (dmlc#5154)

  trivialfis

  authored
  73b1bd2

  Copy full SHA for 73b1bd2

• [Breaking] Remove learning_rates in Python. (dmlc#5155)

  Show description for 298ebe6
  trivialfis

  authored
  298ebe6

  Copy full SHA for 298ebe6

Commits on Dec 26, 2019

• Tests for empty dmatrix. (dmlc#5159)

  trivialfis

  authored
  ced3660

  Copy full SHA for ced3660

• Parameter validation (dmlc#5157)

  Show description for f3d7877
  trivialfis

  authored
  f3d7877

  Copy full SHA for f3d7877

• Enable OpenMP with Apple Clang (Mac default compiler) (dmlc#5146)

  Show description for 9b0af6e
  

  hcho3

  authored and

  trivialfis

  committed
  9b0af6e

  Copy full SHA for 9b0af6e

• implementation of map ranking algorithm on gpu (dmlc#5129)

  Show description for ee81ba8
  

  sriramch

  authored and

  RAMitchell

  committed
  ee81ba8

  Copy full SHA for ee81ba8

Commits on Dec 27, 2019

• Fix wrapping GPU ID and prevent data copying. (dmlc#5160)

  Show description for 61286c6
  trivialfis

  authored
  61286c6

  Copy full SHA for 61286c6

• Clean up Python 2 compatibility code. (dmlc#5161)

  trivialfis

  authored
  6848d04

  Copy full SHA for 6848d04

Commits on Dec 29, 2019

• Don't use modernize-use-trailing-return-type. (dmlc#5169)

  trivialfis

  authored
  d55489a

  Copy full SHA for d55489a

• Fix num_roots to be 1. (dmlc#5165)

  trivialfis

  authored
  139ccc9

  Copy full SHA for 139ccc9

Commits on Dec 30, 2019

• Fix feature_name crated from int64index dataframe. (dmlc#5081)

  

  Hi-king

  authored and

  trivialfis

  committed
  018df60

  Copy full SHA for 018df60

Commits on Dec 31, 2019

• Quick fix for memory leak in CPU Hist. (dmlc#5153)

  Show description for 04db125
  trivialfis

  authored
  04db125

  Copy full SHA for 04db125

• Optimized EvaluateSplut function (dmlc#5138)

  Show description for 7b17e76
  

  SmirnovEgorRu

  authored and

  trivialfis

  committed
  7b17e76

  Copy full SHA for 7b17e76

Commits on Jan 4, 2020

• Fix simple typo: utilty -> utility (dmlc#5182)

  

  timgates42

  authored and

  trivialfis

  committed
  2d95b9a

  Copy full SHA for 2d95b9a

Commits on Jan 7, 2020

• Throw error when not compiled with NCCL. (dmlc#5170)

  trivialfis

  authored
  2b9a62a

  Copy full SHA for 2b9a62a

• Disable parameter validation for Scikit-Learn interface. (dmlc#5167)

  Show description for ebc86a3
  trivialfis

  authored
  ebc86a3

  Copy full SHA for ebc86a3

• Fix span constructor. (dmlc#5166)

  Show description for 77cfbff
  trivialfis

  authored
  77cfbff

  Copy full SHA for 77cfbff

• Lazy initialization of device vector. (dmlc#5173)

  Show description for ee28780
  trivialfis

  authored
  ee28780

  Copy full SHA for ee28780

• Add new lines for Spark XGBoost missing values section (dmlc#5180)

  

  cpfarrell

  authored and

  trivialfis

  committed
  9049c7c

  Copy full SHA for 9049c7c

Commits on Jan 9, 2020

• Move SimpleDMatrix constructor to .cc file (dmlc#5188)

  RAMitchell

  authored
  9559f81

  Copy full SHA for 9559f81

• Implement cudf construction with adapters. (dmlc#5189)

  RAMitchell

  authored
  87ebfc1

  Copy full SHA for 87ebfc1

Commits on Jan 10, 2020

• Remove old cudf constructor code (dmlc#5194)

  RAMitchell

  authored
  8cbcc53

  Copy full SHA for 8cbcc53

Commits on Jan 13, 2020

• Enforce correct data shape. (dmlc#5191)

  Show description for 7b65698
  trivialfis

  authored
  7b65698

  Copy full SHA for 7b65698

• [Breaking] Don't drop trees during DART prediction by default (dmlc#5115)

  Show description for f100b8d
  

  3 peoplecommitted
  f100b8d

  Copy full SHA for f100b8d

• Implement slice via adapters (dmlc#5198)

  RAMitchell

  authored
  a73e25e

  Copy full SHA for a73e25e

Commits on Jan 14, 2020

• Explicitly use UTF-8 codepage when using MSVC (dmlc#5197)

  Show description for 0184f2e
  hcho3

  authored
  0184f2e

  Copy full SHA for 0184f2e

Commits on Jan 15, 2020

• Update R doc by roxygen2. (dmlc#5201)

  trivialfis

  authored
  808f610

  Copy full SHA for 808f610

Commits on Jan 16, 2020

• Fix R dart prediction. (dmlc#5204)

  Show description for 5199b86
  trivialfis

  authored
  5199b86

  Copy full SHA for 5199b86

• Add Optuna badge to README.md (dmlc#5208)

  Show description for e526871
  

  Crissman

  authored and

  trivialfis

  committed
  e526871

  Copy full SHA for e526871

Commits on Jan 18, 2020

• Fix compilation error (dmlc#5215)

  

  mattn

  authored and

  trivialfis

  committed
  ff1342b

  Copy full SHA for ff1342b

Commits on Jan 20, 2020

• Add CMake option to run Undefined Behavior Sanitizer (UBSan) (dmlc#5211)

  Show description for 2a071ce
  

  hcho3

  authored and

  trivialfis

  committed
  2a071ce

  Copy full SHA for 2a071ce

Commits on Jan 22, 2020

• Support dmatrix construction from cupy array (dmlc#5206)

  RAMitchell

  authored
  9c56480

  Copy full SHA for 9c56480

• Fix cupy without cudf import (dmlc#5219)

  RAMitchell

  authored
  5d4c24a

  Copy full SHA for 5d4c24a

• Fix metainfo from DataFrame. (dmlc#5216)

  Show description for 1891cc7
  trivialfis

  authored
  1891cc7

  Copy full SHA for 1891cc7

• working printv C++ command from python

  lav-patel

  committed
  61ece2d

  Copy full SHA for 61ece2d

• working requirements

  lav-patel

  committed
  f752e36

  Copy full SHA for f752e36

• converted choice to string to int

  lav-patel

  committed
  48ddb48

  Copy full SHA for 48ddb48

Commits on Jan 23, 2020

• choice converted to template

  lav-patel

  committed
  d82b45c

  Copy full SHA for d82b45c

• working weighted probabilites

  lav-patel

  committed
  8518701

  Copy full SHA for 8518701

Commits on Feb 7, 2020

• just note

  lav-patel

  committed
  1000f49

  Copy full SHA for 1000f49

Commits on Feb 10, 2020

• oops

  lav-patel

  committed
  f6ff74d

  Copy full SHA for f6ff74d

• testing script

  lav-patel

  committed
  0416633

  Copy full SHA for 0416633

• testing script

  lav-patel

  committed
  86deebb

  Copy full SHA for 86deebb

Commits on Feb 17, 2020

• feb 16

  lav-patel

  committed
  68212dd

  Copy full SHA for 68212dd

• formating

  lav-patel

  committed
  14562d3

  Copy full SHA for 14562d3

• making sure righ xgboost version

  lav-patel

  committed
  b5fb9a8

  Copy full SHA for b5fb9a8

• with sample_weight

  lav-patel

  committed
  c1eb346

  Copy full SHA for c1eb346

• first finding

  lav-patel

  committed
  423d3c9

  Copy full SHA for 423d3c9

• added venv

  lav-patel

  committed
  b2229c9

  Copy full SHA for b2229c9

• working weights

  lav-patel

  committed
  b9acd96

  Copy full SHA for b9acd96

• move files to one level above

  lav-patel

  committed
  4a7590e

  Copy full SHA for 4a7590e

• making sure ownership is lpatel

  lav-patel

  committed
  c7dcd18

  Copy full SHA for c7dcd18

Commits on Feb 20, 2020

• ran the results for weight2

  lav-patel

  committed
  e5faa0c

  Copy full SHA for e5faa0c

Commits on Feb 24, 2020

• requirements.txt

  

  lpatel

  committed
  6109249

  Copy full SHA for 6109249

• updated to run on server

  

  lpatel

  committed
  380bd35

  Copy full SHA for 380bd35

Commits on Mar 3, 2020

• weight1 full grid search

  

  lpatel

  committed
  6c21854

  Copy full SHA for 6c21854

Commits on Mar 4, 2020

• working baysian

  

  lpatel

  committed
  571e16d

  Copy full SHA for 571e16d

Commits on Mar 5, 2020

• after saving first run

  

  lpatel

  committed
  0909b6a

  Copy full SHA for 0909b6a

• w1 baysian with song 's parameter

  

  lpatel

  committed
  bd85cd3

  Copy full SHA for bd85cd3

• jupytext and baysian grid search

  

  lpatel

  committed
  c895605

  Copy full SHA for c895605

• w2

  

  lpatel

  committed
  879ff17

  Copy full SHA for 879ff17

• w3

  

  lpatel

  committed
  3722af7

  Copy full SHA for 3722af7

Commits on Mar 6, 2020

• fixed eta with w1

  

  lpatel

  committed
  ea1d5b5

  Copy full SHA for ea1d5b5

• re w1 oops

  

  lpatel

  committed
  156b971

  Copy full SHA for 156b971

• running parallel

  

  lpatel

  committed
  97726e1

  Copy full SHA for 97726e1

Commits on Mar 9, 2020

• get best models' feature importance

  

  lpatel

  committed
  e19d51e

  Copy full SHA for e19d51e

Commits on Apr 7, 2020

• lpatel to username

  

  lpatel

  committed
  5aa674b

  Copy full SHA for 5aa674b

Commits on Jul 8, 2020

• uncommitted changes from the last time

  

  lpatel

  committed
  13ff0a6

  Copy full SHA for 13ff0a6

Commits on Jul 20, 2020

• featture importance for each iteration

  lav-patel

  committed
  16df369

  Copy full SHA for 16df369

Commits on Aug 4, 2020

• gain feature importance every iteration

  

  lpatel

  committed
  3ec7dff

  Copy full SHA for 3ec7dff

• feature importance by gain every iteration

  

  lpatel

  committed
  6380be6

  Copy full SHA for 6380be6

• previous changes

  

  lpatel

  committed
  5f92a26

  Copy full SHA for 5f92a26

Commits on Aug 5, 2020

• adde bash script

  

  lpatel

  committed
  d5584a6

  Copy full SHA for d5584a6

• fixed typo

  

  lpatel

  committed
  6732116

  Copy full SHA for 6732116

• colsample_bytree_weight pass between python and c as int

  

  lpatel

  committed
  6486666

  Copy full SHA for 6486666

Commits on Aug 6, 2020

• able to pass int vector all the way to random function

  

  lpatel

  committed
  493d0e2

  Copy full SHA for 493d0e2

Commits on Aug 7, 2020

• able to pass multification/devidie int

  

  lpatel

  committed
  31aba1b

  Copy full SHA for 31aba1b

Commits on Aug 8, 2020

• convert float tuple to int

  

  lpatel

  committed
  4338dd2

  Copy full SHA for 4338dd2

• auto formated

  

  lpatel

  committed
  6ba6940

  Copy full SHA for 6ba6940

• convert_vec_to_float compilable

  

  lpatel

  committed
  0cf7f05

  Copy full SHA for 0cf7f05

• static weight(int) works from python

  

  lpatel

  committed
  6d514d5

  Copy full SHA for 6d514d5

Commits on Aug 9, 2020

• model iteration

  

  lpatel

  committed
  ad88ed3

  Copy full SHA for ad88ed3

Commits on Aug 10, 2020

• working iteration and logloss

  

  lpatel

  committed
  2d5403d

  Copy full SHA for 2d5403d

• code clean up

  

  lpatel

  committed
  b9e790a

  Copy full SHA for b9e790a

Commits on Aug 13, 2020

• formatted code

  

  lpatel

  committed
  0859f8c

  Copy full SHA for 0859f8c

Commits on Aug 14, 2020

• refactoring code

  

  lpatel

  committed
  ff30cd6

  Copy full SHA for ff30cd6

• ignore ipython notebook

  

  lpatel

  committed
  594b694

  Copy full SHA for 594b694

• python api

  

  lpatel

  committed
  24bdabc

  Copy full SHA for 24bdabc

Commits on Aug 15, 2020

• find_view_weights

  

  lpatel

  committed
  3a0e13c

  Copy full SHA for 3a0e13c

Commits on Aug 16, 2020

• unused_views_weight

  

  lpatel

  committed
  317ab4e

  Copy full SHA for 317ab4e

Commits on Aug 17, 2020

• new_view_weight_normalized

  

  lpatel

  committed
  36d6bf1

  Copy full SHA for 36d6bf1

• almost done with algo

  

  lpatel

  committed
  d13c8d4

  Copy full SHA for d13c8d4

• autopep8

  

  lpatel

  committed
  38e5fe2

  Copy full SHA for 38e5fe2

• run all weights and see filenames

  

  lpatel

  committed
  0fcf6a8

  Copy full SHA for 0fcf6a8

• ignore nohup.out

  

  lpatel

  committed
  8c6df6b

  Copy full SHA for 8c6df6b

Commits on Aug 23, 2020

• working auc code

  

  lpatel

  committed
  97c9984

  Copy full SHA for 97c9984

• organized function

  

  lpatel

  committed
  d7d74f1

  Copy full SHA for d7d74f1

• running for all weight with 1000 iteration

  

  lpatel

  committed
  76f9bba

  Copy full SHA for 76f9bba

Commits on Aug 24, 2020

• can run w0 to w4

  

  lpatel

  committed
  469d38d

  Copy full SHA for 469d38d

Commits on Aug 25, 2020

• w5 working

  

  lpatel

  committed
  f55aa20

  Copy full SHA for f55aa20

• 500 iteration for all w

  

  lpatel

  committed
  c726cd1

  Copy full SHA for c726cd1

Commits on Aug 27, 2020

• 100 iteration

  

  lpatel

  committed
  1246984

  Copy full SHA for 1246984

Commits on Aug 31, 2020

• commented out repated function

  

  lpatel

  committed
  6ae4c20

  Copy full SHA for 6ae4c20

• working from saving file

  

  lpatel

  committed
  ee8699c

  Copy full SHA for ee8699c

• running on aug30

  

  lpatel

  committed
  e94e496

  Copy full SHA for e94e496

Commits on Sep 15, 2020

• running small scale

  

  lpatel

  committed
  2242546

  Copy full SHA for 2242546

Commits on Sep 16, 2020

• does not work

  

  lpatel

  committed
  7eb29cd

  Copy full SHA for 7eb29cd

• fixed all weigt not totalling to 1

  

  lpatel

  committed
  13031ff

  Copy full SHA for 13031ff

• formating

  

  lpatel

  committed
  058b872

  Copy full SHA for 058b872

• running w1 with 1000000 factor

  

  lpatel

  committed
  ac01dfe

  Copy full SHA for ac01dfe

• running for entiere data set

  

  lpatel

  committed
  bd5d621

  Copy full SHA for bd5d621

• timestamp

  

  lpatel

  committed
  3145ee0

  Copy full SHA for 3145ee0

Commits on Sep 17, 2020

• It stops building tree if it does not see improvement in the last 20 trees(model_iteration).

  

  lpatel

  committed
  8b4a4dd

  Copy full SHA for 8b4a4dd

• w1 and w2 only model_iteration = 500, early_stopping_n = 20

  

  lpatel

  committed
  3ce0537

  Copy full SHA for 3ce0537

Commits on Sep 18, 2020

• printing cv output filename

  

  lpatel

  committed
  9222a1f

  Copy full SHA for 9222a1f

Commits on Sep 19, 2020

• print process

  

  lpatel

  committed
  59824e2

  Copy full SHA for 59824e2

Commits on Sep 22, 2020

• added early stopping at 0.9998 of max of previous 20

  

  lpatel

  committed
  ed489a5

  Copy full SHA for ed489a5

Commits on Sep 23, 2020

• smart conversion from float to int

  

  lpatel

  committed
  d437aa5

  Copy full SHA for d437aa5

Commits on Sep 24, 2020

• 100 model iteration instead of 500 as it is taking lont time

  

  lpatel

  committed
  c1f1477

  Copy full SHA for c1f1477

• profiling comment

  

  lpatel

  committed
  6c69399

  Copy full SHA for 6c69399

Commits on Sep 30, 2020

• effecient choice function

  

  lpatel

  committed
  ea735c5

  Copy full SHA for ea735c5

• commented out method of doing profiling

  

  lpatel

  committed
  d0c53e5

  Copy full SHA for d0c53e5

• added doc string

  

  lpatel

  committed
  e906bb3

  Copy full SHA for e906bb3

Commits on Oct 5, 2020

• installed c++ boost lib

  

  lpatel

  committed
  cb2d363

  Copy full SHA for cb2d363

• faster implementation of choice

  

  lpatel

  committed
  67d3c25

  Copy full SHA for 67d3c25

• choice_n

  

  lpatel

  committed
  e01045a

  Copy full SHA for e01045a

Commits on Oct 6, 2020

• working choice_n function in file

  

  lpatel

  committed
  d38f46b

  Copy full SHA for d38f46b

• compilation is almost done

  

  lpatel

  committed
  0dc70b3

  Copy full SHA for 0dc70b3

• commented out unneeded code

  

  lpatel

  committed
  4be19ed

  Copy full SHA for 4be19ed

• remove unneded code

  

  lpatel

  committed
  35e5069

  Copy full SHA for 35e5069

• renamed the othe same name of the function

  

  lpatel

  committed
  161faba

  Copy full SHA for 161faba

• included choice code in class iteself

  

  lpatel

  committed
  c5595a0

  Copy full SHA for c5595a0

Commits on Oct 7, 2020

• clean up cache privious lib output

  

  lpatel

  committed
  44c88dd

  Copy full SHA for 44c88dd

• debugging statments

  

  lpatel

  committed
  d4611db

  Copy full SHA for d4611db

Commits on Oct 11, 2020

• made shell script logging

  

  lpatel

  committed
  1aae432

  Copy full SHA for 1aae432

• more logging

  

  lpatel

  committed
  6f1db06

  Copy full SHA for 6f1db06

Commits on Oct 12, 2020

• removed print statments

  

  lpatel

  committed
  347b93b

  Copy full SHA for 347b93b

• model_iteration and early_stopping_at at prod level

  

  lpatel

  committed
  39c7448

  Copy full SHA for 39c7448

• removed print stmnt

  

  lpatel

  committed
  cf5bd09

  Copy full SHA for cf5bd09

Commits on Oct 29, 2020

• working with old data set

  

  lpatel

  committed
  f87412f

  Copy full SHA for f87412f

Commits on Nov 3, 2020

• anaconda management

  

  lpatel

  committed
  b83b628

  Copy full SHA for b83b628

Commits on Nov 10, 2020

• trying to read r data

  

  lpatel

  committed
  e93e536

  Copy full SHA for e93e536

• non transpose code

  

  lpatel

  committed
  67ebdf2

  Copy full SHA for 67ebdf2

• notes from meeting

  

  lpatel

  committed
  501e39a

  Copy full SHA for 501e39a

Commits on Jan 25, 2021

• last uncommited changes

  

  lpatel

  committed
  2071286

  Copy full SHA for 2071286

Commits on Mar 16, 2021

• Checked.

  yuanborong

  committed
  f2aac6d

  Copy full SHA for f2aac6d

• Add new random.h .

  yuanborong

  committed
  df237f9

  Copy full SHA for df237f9

Commits on Jun 5, 2021

• test

  yuanborong

  committed
  5a646b7

  Copy full SHA for 5a646b7

Commits on Jun 6, 2021

• add new function (calculateGradientByLogisticClassification(pred, y_train))

  yuanborong

  committed
  dc39c06

  Copy full SHA for dc39c06

• changed xgb.train function

  yuanborong

  committed
  ba8a72e

  Copy full SHA for ba8a72e

• change some bug

  yuanborong

  committed
  172be58

  Copy full SHA for 172be58

• change some bug

  yuanborong

  committed
  b6fe770

  Copy full SHA for b6fe770

• fix bug

  yuanborong

  committed
  97042d7

  Copy full SHA for 97042d7

• change some bug

  yuanborong

  committed
  20ebc87

  Copy full SHA for 20ebc87

• change some bug

  yuanborong

  committed
  026da4d

  Copy full SHA for 026da4d

• test

  yuanborong

  committed
  3cc28b5

  Copy full SHA for 3cc28b5

• test

  yuanborong

  committed
  9c4813c

  Copy full SHA for 9c4813c