Xavier Dupré

It is a known issue. A new rule must be added to the definition of operators TreeEnsembleRegressor and TreeEnsembleClassifier to support that scenario. That's the first step before implementing this...

This was one mayn years. I don't remember the reasons with everything we did. This PR https://github.com/onnx/onnx/pull/5874 changes the way TreeEnsembleRegressor and TreeEnsembleClassifier are defined. It is now one single...

Do you have an example I could use to replicate the issue?

Would it be possible to share an example of what you try to do. I'd like to make it available in sklearn-onnx instead and remove the need for this dependency.

This PR will have a significant impact on converting librairies.

`raw_operator` is the operator coming from scikit-learn. You may assume it is the estimator being converted. The documentation to look at the scikit-learn's.

Not yet. If only the training is different from XGBRgressor, it should not be too difficult to add. Let me investigate.

I created a PR to support these models. It seems to work with the same converter but it could be helpful to check on your side as well.

https://github.com/onnx/onnxmltools/pull/665 (it appears above but maybe that's just for me).

We plan to release a new version before December 14th.