Skip to content

Instantly share code, notes, and snippets.

@kretes

kretes/train.libsvm

Last active March 4, 2019 21:29
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save kretes/1228e571aeba2a57f617352af633cd40 to your computer and use it in GitHub Desktop.
Save kretes/1228e571aeba2a57f617352af633cd40 to your computer and use it in GitHub Desktop.
Download ZIP
xgboost_rank_ndcg_vs_rank_pairwise
Raw
train.libsvm
0 qid:1369666032782981875 0:1.2172619104385376
0 qid:1369666032782981875 0:1.5916666984558105
1 qid:1369666032782981875 0:1.3103448152542114
0 qid:1369666032782981875 0:0.7198443412780762
0 qid:1369666032782981875 0:0.6421052813529968
0 qid:1369666032782981875 0:2.450000047683716
0 qid:1369666032782981875 0:0.3511904776096344
0 qid:1369666032782981875 0:2.110119104385376
0 qid:1369666032782981875 0:1.7380952835083008
0 qid:1369666032782981875 0:1.2692307233810425
0 qid:1369666032782981875 0:1.101190447807312
0 qid:1369666032782981875 0:1.110497236251831
0 qid:1369666032782981875 0:0.75
0 qid:1369666032782981875 0:2.433333396911621
0 qid:1369666032782981875 0:0.7439613342285156
0 qid:1369666032782981875 0:0.5565476417541504
0 qid:1369666032782981875 0:0.6964285969734192
0 qid:1369666032782981875 0:0.720634937286377
0 qid:1369666032782981875 0:0.586309552192688
0 qid:1369666032782981875 0:0.4553571343421936
0 qid:1369666032782981875 0:0.6345177888870239
0 qid:1369666032782981875 0:0.7142857313156128
0 qid:1369666032782981875 0:0.032738097012043
0 qid:1369666032782981875 0:0.4732142984867096
0 qid:1369666032782981875 0:0.4910714328289032
0 qid:1369666032782981875 0:0.4285714328289032
0 qid:1369666032782981875 0:0.449404776096344
0 qid:1369666032782981875 0:0.3988095223903656
0 qid:1369666032782981875 0:0.3452380895614624
0 qid:1369666032782981875 0:0.3839285671710968
0 qid:1369666032782981875 0:0.46000000834465027
0 qid:1369666032782981875 0:0.3482142984867096
0 qid:1369666032782981875 0:0.3214285671710968
0 qid:1369666032782981875 0:0.35748791694641113
0 qid:1369666032782981875 0:0.4166666567325592
0 qid:1369666032782981875 0:0.3511904776096344
0 qid:1369666032782981875 0:0.34408602118492126
0 qid:1369666032782981875 0:0.2708333432674408
0 qid:1369666032782981875 0:0.3392857015132904
0 qid:1369666032782981875 0:0.2589285671710968
0 qid:1369666032782981875 0:0.37735849618911743
0 qid:1369666032782981875 0:1.899999976158142
0 qid:1369666032782981875 0:0.1994047611951828
0 qid:1369666032782981875 0:0.18218623101711273
0 qid:1369666032782981875 0:0.6666666865348816
0 qid:1369666032782981875 0:0.1785714328289032
0 qid:1369666032782981875 0:0.1666666716337204
0 qid:1369666032782981875 0:0.1636904776096344
0 qid:1369666032782981875 0:0.19565217196941376
0 qid:1369666032782981875 0:0.1398809552192688
0 qid:1369666032782981875 0:0.1666666716337204
0 qid:1369666032782981875 0:0.3636363744735718
0 qid:1369666032782981875 0:0.1011904776096344
0 qid:1369666032782981875 0:0.1599999964237213
0 qid:1369666032782981875 0:0.21276596188545227
0 qid:1369666032782981875 0:0.04054053872823715
0 qid:1369666032785346826 0:7.44345235824585
0 qid:1369666032785346826 0:21.053571701049805
0 qid:1369666032785346826 0:6.497023582458496
0 qid:1369666032785346826 0:9.559523582458496
0 qid:1369666032785346826 0:5.80654764175415
0 qid:1369666032785346826 0:6.773809432983398
0 qid:1369666032785346826 0:6.3839287757873535
0 qid:1369666032785346826 0:4.55654764175415
0 qid:1369666032785346826 0:3.169642925262451
0 qid:1369666032785346826 0:5.169642925262451
0 qid:1369666032785346826 0:2.074404716491699
1 qid:1369666032785346826 0:1.753731369972229
0 qid:1369666032785346826 0:4.586309432983398
0 qid:1369666032785346826 0:4.714285850524902
0 qid:1369666032785346826 0:3.5892856121063232
0 qid:1369666032785346826 0:3.657738208770752
0 qid:1369666032785346826 0:2.3333332538604736
0 qid:1369666032785346826 0:3.2771739959716797
0 qid:1369666032785346826 0:3.705357074737549
0 qid:1369666032785346826 0:2.5208332538604736
0 qid:1369666032785346826 0:3.3184523582458496
0 qid:1369666032785346826 0:3.330357074737549
0 qid:1369666032785346826 0:4.083333492279053
0 qid:1369666032785346826 0:4.139881134033203
0 qid:1369666032785346826 0:4.184523582458496
0 qid:1369666032785346826 0:3.732142925262451
0 qid:1369666032785346826 0:2.1190476417541504
0 qid:1369666032785346826 0:4.872023582458496
0 qid:1369666032785346826 0:2.8035714626312256
0 qid:1369666032785346826 0:3.2232143878936768
0 qid:1369666032785346826 0:2.5956521034240723
0 qid:1369666032785346826 0:2.9375
0 qid:1369666032785346826 0:3.6636905670166016
0 qid:1369666032785346826 0:3.809523820877075
0 qid:1369666032785346826 0:2.613095283508301
0 qid:1369666032785346826 0:3.955357074737549
0 qid:1369666032785346826 0:2.824404716491699
0 qid:1369666032785346826 0:1.264880895614624
0 qid:1369666032785346826 0:3.860119104385376
0 qid:1369666032785346826 0:1.6309523582458496
0 qid:1369666032785346826 0:2.267857074737549
0 qid:1369666032785346826 0:3.169642925262451
0 qid:1369666032785346826 0:3.247023820877075
0 qid:1369666032785346826 0:2.9017856121063232
0 qid:1369666032785346826 0:1.6883116960525513
0 qid:1369666032785346826 0:2.9107143878936768
0 qid:1369666032785346826 0:2.529761791229248
0 qid:1369666032785346826 0:2.0731706619262695
0 qid:1369666032785346826 0:1.9940476417541504
0 qid:1369666032785346826 0:2.507042169570923
0 qid:1369666032785346826 0:2.800595283508301
0 qid:1369666032785346826 0:1.4047619104385376
0 qid:1369666032785346826 0:2.4940476417541504
0 qid:1369666032785346826 0:2.4375
0 qid:1369666032785346826 0:2.6011905670166016
0 qid:1369666032785346826 0:1.3777090311050415
0 qid:1369666032785346826 0:2.809523820877075
0 qid:1369666032785346826 0:2.2410714626312256
0 qid:1369666032785346826 0:2.125
0 qid:1369666032785346826 0:2.1875
Raw
xgb_rank.py
import xgboost as xgb
def train(dm, objective):
model = xgb.train(params={
"max_depth": 6,
"subsample": 1,
"colsample_bytree": 1,
"objective": objective,
"eval_metric": ["ndcg@60"],
"silent": 1,
"seed": 42,
"booster": "gbtree",
"tree_method": "hist"
},dtrain=dm, num_boost_round=10,
evals=[(dm,"train")],
verbose_eval=True,
maximize=True)
dm = xgb.DMatrix("train.libsvm")
train(dm, "rank:pairwise")
train(dm, "rank:ndcg")
Raw
xgboost_rank_ndcg_vs_pairwise
[22:25:37] 116x1 matrix with 116 entries loaded from /tmp/116
[22:25:37] Tree method is selected to be 'hist', which uses a single updater grow_fast_histmaker.
[0] train-ndcg@60:0.815465
[1] train-ndcg@60:1
[2] train-ndcg@60:1
[3] train-ndcg@60:1
[4] train-ndcg@60:1
[5] train-ndcg@60:1
[6] train-ndcg@60:1
[7] train-ndcg@60:1
[8] train-ndcg@60:1
[9] train-ndcg@60:1
[22:25:38] Tree method is selected to be 'hist', which uses a single updater grow_fast_histmaker.
[0] train-ndcg@60:0.815465
[1] train-ndcg@60:0.815465
[2] train-ndcg@60:0.815465
[3] train-ndcg@60:0.815465
[4] train-ndcg@60:0.815465
[5] train-ndcg@60:0.815465
[6] train-ndcg@60:0.815465
[7] train-ndcg@60:0.815465
[8] train-ndcg@60:0.815465
[9] train-ndcg@60:0.815465
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment