If we specify both "EmbeddingIntentClassifier" and "SklearnIntentClassifier" in the NLU pipeline configuration, will Rasa do some kind of ensemble or only one of the classifiers will be used for intent classification?
Hi @saurabh-m523, an ensemble model is not created. Both the classifiers will be trained during training and both will process the message during inferencing. But depending on the order in which you specify them in the NLU config file, you’ll be able to see the output of only one(the one declared second). The second classifier will overwrite the output of the first during inference. Hence, it is recommended to use only one classifier in your NLU pipeline
which one is good?