LightGBMError: b'Wrong size of feature_names'

Created on 16 Nov 2017  路  4Comments  路  Source: microsoft/LightGBM

Hello there, when i run my dataset with lgb, following errors occur:

_Data: X_train: (314954, 76), X_test: (78739, 76)
Training on: (314954, 76)
Traceback (most recent call last):

File "", line 80, in
model = lgb.train(params, train_lgb, num_boost_round=ROUNDS)

File "C:\Users\User\Anaconda3\lib\site-packages\lightgbm\engine.py", line 213, in train
booster._load_model_from_string(booster._save_model_to_string())

File "C:\Users\User\Anaconda3\lib\site-packages\lightgbm\basic.py", line 1554, in _load_model_from_string
ctypes.byref(self.handle)))

File "C:\Users\User\Anaconda3\lib\site-packages\lightgbm\basic.py", line 48, in _safe_call
raise LightGBMError(_LIB.LGBM_GetLastError())

LightGBMError: b'Wrong size of feature_names'_

Does somebody know how to deal with this issues? thanks.

All 4 comments

ping @wxchan @StrikerRUS

@ahbon123 Please provide reproducible example or at least the params and the head of train_lgb.

@ahbon123 what's the type of X_train and X_test? and what's the value of ROUNDS?

Thanks for your reply @StrikerRUS @wxchan @guolinke. The bugs come from the features' titles. When I delete the row of titles, there is no more errors.

Was this page helpful?
0 / 5 - 0 ratings