Here we wish to use some of the MASSIVE dataset from Amazon used to train and
evaluate virtual assistants. We use only the data in English, found under data/en-US.jsonl
and filter it to just a
small number of intents in the notebook generate_chatbot_data.ipynb
. The result is essentially the training data
found in data/nlu.yml
. To run:
- clone the repo with
git clone [email protected]:clharris/massive_rasa.git
- run
conda env create --file=environment.yml
- run
conda activate massive_rasa
- run
rasa train
to train a model - run
rasa interactive
for an interactive session with the chatbot
See more details in the rasa docs.