H2O is an Open Source, Distributed, Fast & Scalable Machine Learning Platform: Deep Learning, Gradient Boosting (GBM) & XGBoost, Random Forest, Generalized Linear Modeling (GLM with Elastic Net), K-Means, PCA, Generalized Additive Models (GAM), RuleFit, Support Vector Machine (SVM), Stacked Ensembles, Automatic Machine Learning (AutoML), etc.
But I am in doubt now, I think maybe the problem is H2O itself.
When I do a {{aml_lead = aml.leader.download_mojo(path="/tmp/")}} the output is the same as if I issue {{aml_lead = aml.download_mojo(path="/tmp/")}}. And I can not see the binaries in the resulting zip file.
If anyone, for example, exports a trained GBM model, it would be exported as an ONNX successfully. Inside the zip, you would find the .bin files inside the /trees/ folder
But I am unable to export a model which has previously been trained through AutoML. If I try to export the Ensemble, XGBoost or GLM models, it will not work.
I initially posted the bug here: [https://github.com/onnx/onnxmltools/issues/409|https://github.com/onnx/onnxmltools/issues/409]
But I am in doubt now, I think maybe the problem is H2O itself.
When I do a {{aml_lead = aml.leader.download_mojo(path="/tmp/")}} the output is the same as if I issue {{aml_lead = aml.download_mojo(path="/tmp/")}}. And I can not see the binaries in the resulting zip file.
If anyone, for example, exports a trained GBM model, it would be exported as an ONNX successfully. Inside the zip, you would find the .bin files inside the /trees/ folder But I am unable to export a model which has previously been trained through AutoML. If I try to export the Ensemble, XGBoost or GLM models, it will not work.