Parse a message with the trained nlu model

Rasa Open Source
1

using the last version of rasa 1, after training the model, I’m trying to run the nlu server to parse sentence with it using the command: python -m rasa_nlu.server --path [model] so after the server runs, when I pass a message to it as follows:

curl 'http://localhost:5000/parse?q=terrible'

what i always gets as a result is:

{ “intent”: { “name”: null, “confidence”: 1.0 }, “entities”: [], “text”: “terrible”, “project”: “default”, “model”: “fallback”

it seems like I’m not actually passing any trained models however I did that already with the command above while starting the nlu server any help please?

2

We changed the command line interface in Rasa 1.0. Are you sure you are using Rasa 1.x? The command python -m rasa_nlu.server --path [model] should not work with Rasa 1.x. We combined the Rasa NLU and Core server. There is only one server now. To parse a message with latest Rasa version, you should execute the following steps:

  1. Train a Rasa model using rasa train. If you just want to train an NLU model use rasa train nlu.
  2. Start a Rasa server with rasa run --enable-api
  3. Parse a message by running
curl -X POST \
  http://localhost:5005/model/parse \
  -d '{
    "text": "Hello!"
}'

Please, check out the documentation on the Command Line Interface for more information.

1 Like