Save LGBMRegressor model from python lightgbm package to disc

python lightgbm
Utpal Datta picture Utpal Datta · Mar 17, 2019 · Viewed 16.5k times · Source

Hi I am unable to find an way to save and reuse an LGBM model to a file. I used python package lightgbm and LGBMRegressor model. Could you please help? Documentations doesn't seem to have useful information. I am using python 3.5 on Spyder

Answer

Prayson W. Daniel picture Prayson W. Daniel · Mar 17, 2019

Try:

my_model.booster_.save_model('mode.txt')
#load from model:

bst = lgb.Booster(model_file='mode.txt')

You can also use pickle:

import joblib
# save model
joblib.dump(my_model, 'lgb.pkl')
# load model
gbm_pickle = joblib.load('lgb.pkl')

Let me know if that helps

Related questions

Multiclass Classification with LightGBM

I am trying to model a classifier for a multi-class Classification problem (3 Classes) using LightGBM in Python. I used the following parameters. params = {'task': 'train', 'boosting_type': 'gbdt', 'objective': 'multiclass', 'num_class':3, 'metric': 'multi_logloss', 'learning_rate': 0.002296, 'max_depth': 7, 'num_…

Read More
Why ImportError: No module named lightgbm

My OS is Ubuntu, and I've followed the official installation guide to install lightgbm. However, when I import it, this error is raised: ImportError: No module named lightgbm How can I solve this? Do I also need to go to /…

Read More
Feature importance using lightgbm

I am trying to run my lightgbm for feature selection as below; initialization # Initialize an empty array to hold feature importances feature_importances = np.zeros(features_sample.shape[1]) # Create the model with several hyperparameters model = lgb.LGBMClassifier(objective='binary', boosting_…

Read More