Ep3 Masterclass - changing the pipeline

Getting Started with Rasa
1

I have created a new bot called MedicareLocator following the instructions but when I look at the config.yml file, it is not identical to that shown in the video/e-book. Mine looks like this:

image
image1254×965 42.4 KB
When I add in supervised-embeddings like this:
image
image1487×949 44.9 KB
I get this error:
image
image1902×199 11.9 KB
What am I doing wrong?
TIA

2

The masterclass videos are a bit old now. They were using Rasa 1.x while we just released Rasa 2.6. I fear the config file you have is using a shortcut which is now deprecated.

Instead, I would recommend using the config.yml file that you get when you run rasa init. I’ve copied the base config below just in case.

language: en

pipeline:
 - name: WhitespaceTokenizer
 - name: RegexFeaturizer
 - name: LexicalSyntacticFeaturizer
 - name: CountVectorsFeaturizer
 - name: CountVectorsFeaturizer
   analyzer: char_wb
   min_ngram: 2
   max_ngram: 4
 - name: DIETClassifier
   epochs: 100
   constrain_similarities: true
   model_confidence: linear_norm

policies:
 - name: MemoizationPolicy
 - name: TEDPolicy
   max_history: 5
   epochs: 50
   constrain_similarities: true
 - name: RulePolicy
3

Thank you very much for your help!

What is the recommended learning path now? Or could I carry on with the Masterclass but look the current syntax up in some documentation?

BW Lorna

4

We’re working on a new learning path as we speak but in the meantime you might appreciate this blogpost on the machine learning pipeline.There’s also this youtube playlist with small tutorials.

5

That’s fantastic. Many thanks!