Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

Can i train mml-lgb model and loading in local Python?

See original GitHub issue

Code like below

from mmlspark.lightgbm import LightGBMClassifier
model =LightGBMClassifier(boostingType='gbdt',
                          objective='binary',
                          baggingFreq=5,
                          learningRate=0.1,
                          numIterations=50,
                          earlyStoppingRound=10,
                          featuresCol = 'features',
                          categoricalSlotNames = cate_list,
                          numLeaves=30,
                          labelCol="label").fit(train)

model.saveNativeModel(path)

from sklearn.externals import joblib
clf = joblib.load(path)

Issue Analytics

State:
Created 2 years ago
Comments:9 (5 by maintainers)

Top GitHub Comments

1reaction

imatiach-msftcommented, Jun 14, 2021

you can create the python booster from model str: https://github.com/microsoft/LightGBM/issues/2097 model = lgb.Booster({‘model_str’: buf.read().decode(“UTF-8”)}) you can also create the scikit-learn API but it’s not “officially supported”: https://github.com/microsoft/LightGBM/issues/1942

0reactions

RainFungcommented, Jun 17, 2021

Thanks，I finished the project alreday.