Closed vruusmann closed 5 years ago
Related to https://github.com/jpmml/jpmml-evaluator/issues/96
In other words, there should be an API for producing transformer-only PMML documents, and then consuming them.
Hello, When could we expect this "transformer-only" conversion mode? And our team is looking forward to this mode :) Moreover, if it is possible, a full example could be provided first before this mode, e.g for PCA. Because I am not familiar with PMML file, it is difficult and confusing for me to adjust relevant configuration as your suggestion in Support for transformer-only pipelines. Thank you!
Hello, When could we expect this "transformer-only" conversion mode? And my team is looking forward to this mode :) I am not familiar with PMML file, it is difficult and confusing for me to adjust relevant configuration as your suggestion in Support for transformer-only pipelines. And how do I look at the results of the data transformation, which is what the data looks like before I put it into the evaluator? Thank you!
Based on https://github.com/jpmml/jpmml-sklearn/issues/86
At the moment it's impossible to generate transformer-only pipelines, because the
ModelEncoder#encodePMML(Model)
method applies a set of visitors that clean the soon-to-be-generated PMML document from all unused preprocessing instructions: https://github.com/jpmml/jpmml-converter/blob/master/src/main/java/org/jpmml/converter/ModelEncoder.java#L53-L56Possible solution: class
ModelEncoder
should provide a "transformer-only" conversion mode.