Tensorflow pipeline training never ends

Rasa Open Source
1

Hello all,

I am new to Rasa Nlu, my story is when I use tensorflow pipeline to train, if only 1 intent (below example) then it will finish fast (only comes with message ERROR rasa_nlu.classifiers.embedding_intent_classifier - Can not train an intent classifier. Need at least 2 different classes. Skipping training of intent classifier.

=============================================================

intent: meetup

  • I am new to the area. What meetups I could join in Berlin?
  • I have just moved to Berlin. Can you suggest any cool meetups for me?
  • I am live in Taiwan, can we meet? =============================================================

but when I have have 1+ intents (not talking about muti intents in one sentencce) as below, the train will never end.

=============================================================

intent: meetup

  • I am new to the area. What meetups I could join in Berlin?
  • I have just moved to Berlin. Can you suggest any cool meetups for me?
  • I am live in Shatin, can we meet?

intent: meetup1

  • again testing, will it work? =============================================================

can anyone helps, I am stuck for a week already, thanks so much!

  • python 3.6 in ananconda:
  • tensorflow==1.8.0 (1.12.0) also the same
  • ‘rasa-core==0.9.6’ ‘rasa-nlu==0.12.3’

training commend: !python -m rasa_nlu.train -c config_tenosorflow_1.yml --data data/nlu_data1.md -o models --fixed_model_name intentClassifier --project current --verbose

Regards, Ed

2

Hi Ed,

training a classifier with only one class is not very useful so let’s address the other issue.

Are there any errors or does it just train forever?

A good way to find the source of the problem is to start by reproducing an example from the docs and then, piece by piece, transforming it to what you want. See if you can reproduce this model from the docs and then change the pipeline/training data to the one you wish to use

Cheers, Nikolai