Rasa Trained Model is giving incorrect INTENT identifications

Rasa Open Source
1

Hi Everyone, We are using Rasa Open Source in our chatbot application. We are calling Rasa Train API to train the Model. We can able to train the model successfully using API. But the problem is, Rasa is not giving correct intents until we restart the Rasa Server after training is completed through an API.

When we Train the Rasa Model in the Command Prompt, we are not facing this issue. But in real-time, we cannot use the command prompt right. So we are using the APIs to train the Rasa Model. But using API, we have to restart the server after training is completed. Then only we are getting the correct Intent and Entity identification from Rasa

Could anyone please help me with this?

2

Hi Sunil,

Could you text here how do you train your model using RASA API rather than cmd line ?

3

Hi @santhosh-sp Through an API we are running rasa train nlu command in the command prompt and we were able to train the model successfully.

We realized that after training is completed, we have to load the latest model. To do that I am using the below API. but facing issues while doing that. Please check the attached screenshot for your reference.

image
image1525×821 58.2 KB