I have a rasa nlu model running to detect intents and entities. Whenever a new model is loaded by rasa after training, I want to know how good or bad it is from previous model. What I am planning is , whenever a model is loaded first time, I can test it against some gold standard data and save the output metrics to compare with previous saved metrics. Questions: Is there any event which is loaded first and one time …

I want to add few custom attributes with respect to every intent in rasa nlu.yml, but could not find such thing in documentation. Purpose is to further classify intent based on say for a domain or customer. For example, bookappointment, searchreservations are part of Travel domain, while makepayment, addcredit are of Finance Domain. How can I maintain such information for respective intents? If not in nlu.yml then which other can be a better place to maintain it?

As per RASA DietClassifer architecture, it uses transformer before CRF. Also, it optionally can use SpacyNLP. Despite such a complex model with so many layers the execution speed of RASA is very good. Transformers are known for their heavy weight and slowness. Then how is RASA DietClassifer so fast?