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.

Combine sklearn pipeline with keras or lightgbm

See original GitHub issue

Hello! It’s possible to use pipeline for preparing data and on top of that keras or lightgbm model? It is not clear for me - which convert function I should choose for this operation

Issue Analytics

State:
Created 5 years ago
Comments:7

Top GitHub Comments

2reactions

PatrickSagercommented, Oct 23, 2019

@xadupre: I was also not able to export a pipeline that contains a scikit_learn.KerasClassifier. I’m getting the following error: Unable to find a shape calculator for type ‘<class ‘keras.wrappers.scikit_learn.KerasClassifier’>’. Could you please provide an example or add a statement, that this is the scikit_learn.KerasClassifier is currently not supported.

2reactions

xaduprecommented, Jan 31, 2019

We are currently refactoring onnxmltools and keras converters. The mechanism will be same (you need to register a new converter) but it should be more simple. The custom shape calculator should disappear to be replaced by a version coming from onnxmltools. So this line will remain:

update_registered_converter(LGBMClassifier, ‘LightGbmLGBMClassifier’,
lightgbm_classifier_shape_extractor, convert_lightgbm)

But you won’t have to implement lightgbm_classifier_shape_extractor.

Top Results From Across the Web

Using keras models with scikit-learn pipelines - Kaggle

In this code, we will perform cross validation on the Keras model accuracy using the StatifiedKFold method in the sklearn library.

Transfer Learning with ONNX

It is not quite easy in this case to mix framework, scikit-learn with pytorch (or skorch), the Keras API for Tensorflow, tf.keras.wrappers.scikit_learn.

Gradient Boosting with Scikit-Learn, XGBoost, LightGBM, and ...

Gradient boosting refers to a class of ensemble machine learning algorithms that can be used for classification or regression predictive ...

How to build pipeline with grid search cv and early stopping ...

Here is my code. It start with pipeline (standardizing,replace null value,onehotencoding and selectkbest) with lightgbm model to fit my data ...

Extracting Feature Importances from Scikit-Learn Pipelines

Scikit-learn pipelines provide a really simple way to chain together the preprocessing steps with the model fitting stages in machine ...