Storing Rasa NLU models in cloud?

Getting Started with Rasa
1

Hi I am having trouble with storing the models in the aws. Actually I am having trouble writing the proper structure for config file for a particular extention i.e config.yml. for AWS credentials list

  • AWS_SECRET_ACCESS_KEY
  • AWS_ACCESS_KEY_ID
  • AWS_DEFAULT_REGION
  • BUCKET_NAME
  • AWS_ENDPOINT_URL

my config file structure is: pipeline:

  • name: “nlp_spacy”
  • name: “tokenizer_spacy”
  • name: “ner_duckling_http” url: “http://localhost:8000” dimensions: [“time”, “phone-number”, “url”, “email”]
  • name: “ner_crf”
  • name: “ner_synonyms”
  • name: “intent_featurizer_count_vectors”
  • name: “intent_classifier_tensorflow_embedding” intent_tokenization_flag: true intent_split_symbol: “,”

I dont know how to add above aws list to the give structure. I tried different ways. Nothing worked.

2

Hi @sandy130794, these should be set as environment variables, and are not part of the training config.yml. Here’s how you set a variable in your shell (assuming you use bash for example):

export AWS_SECRET_ACCESS_KEY=my_secret_access_key