I trained the rasa framework and got two different models. accuracy of both models is the same. But one model’s button functionality is working properly but the other model’s button functionality is not working well. Could someone pls clarify what could be the reason for this?
Are there any difference in the training data and could you please elaborate on how they behave differently? Also worth sharing your model configuration.
Which Rasa Open Source version are you using? E.g. the component
intent_classifier_tensorflow_embedding doesn’t existing anymore.
The following versions I am using. Rasa-Core = 0.12.0 Rasa-Core-SDK = 0.12.1 Rasa -NLU = 0.13.8
these versions are deprecated since nearly 2 years now. I’d suggest updating to Rasa 2