Hi,
We are trying our hand at Rasa as our NLP. Our application have multiple chatbots managing different functionality. Currently we are managing different NLPs for each. As Rasa is based on models, we are confused how we can handle multiple models in same Rasa instance.
Can we create separate models for each app in same Rasa instance and parse the model according to each chatbot.
And also can we customize the model name while creating it using the model/train API.
Thanks
Anjaly
As Rasa is based on models, we are confused how we can handle multiple models in same Rasa instance.
You can’t have multiple models simultaneously in the same instance. You could load and unload models but this wouldn’t perform well.
Can we create separate models for each app in same Rasa instance and parse the model according to each chatbot.
No
And also can we customize the model name while creating it using the model/train API.
It doesn’t look like this is an option. You can rename it after training or add this as a feature.
Thanks @stephens for the reply.
That means we need to create separate rasa instance for each applications right.
Also we could see --fixed-model-name for custom model name in rasa train command. Can we use the same parameter in model/train API. Or do we have any API for renaming the model
That means we need to create separate rasa instance for each applications right.
You need a separate instance for each model.
Also we could see --fixed-model-name for custom model name in rasa train command.
Yes, in the CLI you have this option. Doesn’t look like it’s in the API though. You could add this feature yourself. Probably here.