KeyError: 'HFTransformersNLP'

Rasa Open Source
1

Hi Team,

I am trying to use BERT in my DIET intent classifier. The versions are: rasa=2.6.2 rasa_core=0.15.1 rasa_nlu=0.14.5 spacy=2.2.4

But it is not identifying “HFTransformersNLP”.

I also tried using LanguageModelFeaturizer only but then i get KeyError: “LanguageModelFeaturizer”

I am trying to run intent classifier in jupyter notebook config2 = “”" language: “en”

pipeline:

  • name: “HFTransformersNLP” model_weights: “bert-base-uncased” model_name: “bert”
  • name: “LanguageModelTokenizer” # splits the sentence into tokens
  • name: “LanguageModelFeaturizer”

model_name: “bert”

model_weights: “rasa/LaBSE”# uses the pretrained Featurizer

  • name: “CountVectorizerFeaturizer” # transform the sentence into a vector representation
  • name: “DIETClassifier”

“”" with open("./config2.yml",“w”) as report: report.write(config2) report.close()

training_data = load_data(“nlu.md”)

trainer to educate our pipeline

trainer2 = Trainer(config.load("./config2.yml"))

train the model!

interpreter2 = trainer2.train(training_data)

store it for future use

model_directory = trainer.persist("./models/nlu2", fixed_model_name=“current2”)

2

Welcome to the community :slight_smile:

To properly format code, please use three backticks (```) a line before and a line after the code to make it readable.

```
like this
```

Please send your pipeline again while doing this, as well the output of the command rasa --version.