I’m confused about how interactive learning works.
I’ve gone through a few rounds of interactive learning with my bot.
I noticed that the stories.md and nlu.md files are updated with my conversations but the model file is not updated i.e. there aren’t any new models.
Am I suppose to manually run the training process to incorporate the new conversations into the model or are the new conversations actually incorporated into the existing model in real-time without changing the model’s name?
the idea is that the models don’t change, they’re versioned, so that if something goes wrong, you can go back to the one that worked. If we were creating a new model every time something changed, that would be a lot of models! So just re-train one when you’re ready to test out the changes you’ve added to your training data.