search

microsoft / LightGBM

A fast, distributed, high performance gradient boosting (GBT, GBDT, GBRT, GBM or MART) framework based on decision tree algorithms, used for ranking, classification and many other machine learning tasks.
https://lightgbm.readthedocs.io/en/latest/
MIT License
16.52k stars 3.82k forks source link

Fatal Error: Tree model string format error #445

Closed Soledad89 closed 7 years ago

Soledad89 commented 7 years ago

Has anybody encountered this error: Met Exceptions: Tree model string format error

I found that the model (LightGBM_model.txt) I have trained contains some codes as below:

num_leaves=1 split_feature= split_gain= threshold= decision_type= left_child= right_child= leaf_parent=-1 leaf_value=0 leaf_count=0 internal_value= internal_count= shrinkage=1

That's the bad boy. How to solve this kind of problem during training process? Thanks so much.

guolinke commented 7 years ago

How do you train this model? python/R or cli? can you paste the full model file?

Soledad89 commented 7 years ago

@guolinke I just run the multiclass_class example with my own datasets, that has 18 classes and 50-dimention feature vector

# task type, support train and predict
task = train

# boosting type, support gbdt for now, alias: boosting, boost
boosting_type = gbdt

# application type, support following application
# regression , regression task
# binary , binary classification task
# lambdarank , lambdarank task
# multiclass
# alias: application, app
objective = multiclass

# eval metrics, support multi metric, delimite by ',' , support following metrics
# l1 
# l2 , default metric for regression
# ndcg , default metric for lambdarank
# auc 
# binary_logloss , default metric for binary
# binary_error
# multi_logloss
# multi_error
metric = multi_logloss

# number of class, for multiclass classification
num_class = 18

# frequence for metric output
metric_freq = 1

# true if need output metric for training data, alias: tranining_metric, train_metric
is_training_metric = true

# number of bins for feature bucket, 255 is a recommend setting, it can save memories, and also has good accuracy. 
max_bin = 255

# training data
# if exsting weight file, should name to "regression.train.weight"
# alias: train_data, train
data = multiclass.train

# valid data
valid_data = multiclass.test

# round for early stopping
early_stopping = 32

# number of trees(iterations), alias: num_tree, num_iteration, num_iterations, num_round, num_rounds
num_trees = 6400

# shrinkage rate , alias: shrinkage_rate
learning_rate = 0.05

# number of leaves for one tree, alias: num_leaf
num_leaves = 128
chivee commented 7 years ago

@Soledad89 , could you also share your training samples snippet?

Soledad89 commented 7 years ago

@chivee Of course yes.

2   -0.998034765141 -0.997706683756 -0.997314689194 -0.9981045614   -0.997629411465 -0.997813119594 -0.999393269667 -0.999013981692 -0.999197535788 -0.999796847852 -0.973528427146 -0.972987702038 -0.972251760183 -0.986813828764 -0.985929265736 -0.989343602974 -0.999619952868 -0.999139498859 -0.998952794858 -0.999572947562 -0.325514941145 0.153019787298  0.42280849477   0.581581249155  0.633799891812  0.521182652645  -0.270423705964 -0.205594218392 -0.587916320576 -0.990597605012 0.0421030110615 0.482631193353  0.774933478788  0.850205908371  0.935949581982  0.950049092198  -0.256683320968 -0.028672272683 -0.287016060665 -0.902583934781 -0.96366088554  -0.810406279636 -0.702659786771 -0.212641054104 -0.339831867621 -0.726196570603 -0.202525470066 -0.331472015525 -0.876332185002 -0.964340924203
3   -0.968713194268 -0.967523066378 -0.966454432214 -0.982745321757 -0.981869484474 -0.987012732186 -0.999578234816 -0.99926941828  -0.999354599343 -0.999813132863 -0.419471484792 -0.418186816554 -0.417099974939 -0.704764290359 -0.703965172059 -0.802370795023 -0.999993085125 -0.999987987943 -0.999988885474 -0.999996916191 -0.0443317570158    0.34915438745   0.655670499293  0.576077969 0.809692729985  0.870883643859  -0.490554318666 -0.172811882427 -0.220984559223 -0.766290715764 -0.0158620682162    0.370899761232  0.675948034777  0.701242095002  0.806455334871  0.710285625362  -0.147313286541 0.343206519509  0.294295055806  -0.575232762198 -0.982143442627 -0.982212203345 -0.976887006792 -0.994005487711 -0.977772241858 -0.951751530691 -0.997804877283 -0.98040291421  -0.955115621008 -0.965338437282
6   -0.992591348142 -0.976423897467 -0.950090779449 -0.949054225517 -0.909602884616 -0.883030880974 -0.969465093935 -0.937088288507 -0.932012070455 -0.973846948348 -0.991161678347 -0.9900767078   -0.985411363793 -0.993544986283 -0.98492197848  -0.973494762898 -0.999127775975 -0.993299849009 -0.983736599179 -0.987928807118 -0.818107519384 -0.816720444336 -0.81573837845  -0.903293636913 -0.902758888102 -0.93094573081  -0.999979924747 -0.99997087571  -0.999978929927 -0.999996453539 0.0491221588525 0.558536892529  0.920938458589  0.961839106659  0.968182427178  0.787485930102  -0.0327539647831    0.460598300324  0.394319042074  -0.5543609606   -0.963686673919 -0.879718841947 -0.831522800027 -0.537922722477 -0.632968397997 -0.87435548232  -0.532848651084 -0.62974107448  -0.947658405355 -0.971689871776
8   -0.962706421755 -0.962513197145 -0.962296953644 -0.981752413113 -0.981580390215 -0.987679757169 -0.999998405072 -0.999996186055 -0.999995271706 -0.99999788899  -0.152433439985 0.352271295228  0.722126006756  0.700069112142  0.923383019876  0.96722741512   -0.188065017871 0.238969751194  0.190109897713  -0.615786028593 -0.71776622776  -0.666520442654 -0.63514392976  -0.742572022475 -0.728497197389 -0.790685447625 -0.98183817228  -0.977514795468 -0.987267428152 -0.999015906566 -0.996829802544 -0.996766779992 -0.996218726896 -0.997958267809 -0.997173040157 -0.996913063201 -0.999996653903 -0.999872360414 -0.999624013563 -0.999666547144 -0.98422694957  -0.793063735152 -0.414908593023 -0.218480435487 0.165259865854  0.198201701912  -0.247029000159 0.137568128167  -0.0518752688574    -0.714254038228
12  -0.962266462331 -0.942764084878 -0.922728625326 -0.912384878866 -0.893404062737 -0.905811711623 -0.945220555013 -0.928130095151 -0.957321388086 -0.994702852455 -0.132818141918 0.416456673144  0.803224087495  0.801892627724  0.987068397583  0.986999828463  -0.0800331140152    0.388993566813  0.360827858554  -0.530095769727 -0.0250430166354    0.484732759447  0.830290918155  0.844200373005  0.973556182567  0.943493480796  -0.106179899405 0.325304939463  0.245590684513  -0.620360230463 0.000135971872172   0.518764687011  0.904767990138  0.95137010251   0.976128510427  0.800280154061  -0.0313981964597    0.477465566468  0.426085672833  -0.525847147235 -0.985689553687 -0.985688783148 -0.989722223799 -0.996564062331 -0.990682530181 -0.956583586166 -0.999999999455 -0.986992218888 -0.94831816322  -0.948321875174
16  -0.881202518069 -0.881139800621 -0.881064142816 -0.941808601346 -0.941718332212 -0.9615308626   -0.999759998541 -0.999670916784 -0.999791801368 -0.999968902259 -0.0812150928462    0.475255105033  0.843309047391  0.85141903908   0.988519765379  0.94708030913   -0.0111169506532    0.447402110696  0.395307611627  -0.538536922857 -0.0816065797724    0.386973942148  0.739443556136  0.725101093138  0.928908362987  0.943754336459  -0.233315135422 0.220016411861  0.200457759368  -0.588040259657 -0.963013165037 -0.959671773946 -0.956650896347 -0.969348697798 -0.967086985472 -0.972292995294 -0.997951268319 -0.996990134236 -0.997839687482 -0.999656856208 -0.982164461763 -0.980760604962 -0.988066755798 -0.993804520598 -0.993968766153 -0.963167657661 -0.999641226836 -0.990770824123 -0.952125878337 -0.947488045946
6   -0.975684293947 -0.972009887276 -0.964915540651 -0.981735556567 -0.973486213297 -0.971440534382 -0.999080400061 -0.996384792762 -0.993589944072 -0.996061831999 -0.981717883481 -0.973630882427 -0.960397645    -0.954674042125 -0.934132675691 -0.927763080884 -0.967709114544 -0.948309046848 -0.956488905399 -0.989785900123 -0.866673877407 -0.862147869908 -0.858105118343 -0.923032105149 -0.920111158209 -0.940570040721 -0.999650356661 -0.999386555526 -0.999434820662 -0.9998443271   -0.549984766385 -0.337376740055 -0.07508887081  -0.0497993811631    0.229497700835  0.339549745673  -0.561209681599 -0.25026300008  -0.29439611272  -0.789745960295 -0.963633650988 -0.864325409777 -0.738249289141 -0.477702621076 -0.427347073918 -0.61531494035  -0.475493050375 -0.424410603465 -0.754537113526 -0.993460399367
10  -0.417282636332 -0.0453515142183    0.280242982318  0.214436254845  0.509130540557  0.577537750385  -0.441779691965 -0.09648992617  -0.162869366956 -0.723504696229 -0.430711101051 0.0135580047269 0.377525214077  0.390467316695  0.648527726564  0.708882348143  -0.350213045977 -0.048169414112 -0.174151271138 -0.795252096024 -0.297967657086 0.168317998929  0.560196942946  0.592452616183  0.844755458606  0.90449468816   -0.206150481702 0.22047319995
guolinke commented 7 years ago

@Soledad89 I mean the LightGBM_model.txt file.

guolinke commented 7 years ago

@Soledad89 BTW, it seems you didn't use the latest code. Can you try the latest code(master branch) ?

Soledad89 commented 7 years ago

@guolinke @chivee Great. While, I have tried the latest code and solved the issue. I suppose that it's my fault. The label should start from zero. That's the bad boy. Thanks soooo much for your kind reply.

github-actions[bot] commented 1 year ago

This issue has been automatically locked since there has not been any recent activity since it was closed. To start a new related discussion, open a new issue at https://github.com/microsoft/LightGBM/issues including a reference to this.