/NLU-training-data

Crowd sourced training data for Rasa NLU models

Apache License 2.0Apache-2.0

NLU Training Data

Crowd-sourced training data for the development and testing of Rasa NLU models.

About this repository

This is an experiment with the goal of providing basic training data for developing chatbots, therefore, this repository is open for contributions!

We need your help to create an open source dataset to empower chatbot makers and conversational AI enthusiasts alike, and we very much appreciate your support in expanding the collection of data available to the community.

How do I donate my training data?

Each folder should contain a list of multiple intents, consider if the set of training data you're contributing could fit within an existing folder before creating a new one.

To contribute via pull request, follow these steps:

  1. Create an issue describing the training data you would like to contribute.

  2. Create a new file with a folder title and a NLU.md file, or contribute to an existing folder.

  3. In the NLU.md file, format your training data using markdown, remove all entities (see script), title each section with the intent types and add a short description e.g.intent:inform_rain <!--The user says that it is currently raining somewhere.-->

  4. Update the README.md file, include a list of the intent types added.

  5. Create a pull request describing your changes.

Your pull request will be reviewed by a maintainer, who will get back to you about any necessary changes or questions. You will also be asked to sign a Contributor License Agreement.

FAQs

How should I label my intents?

Please always put the domain at the end of each intent. For example: ask_transport

What do I do about multi-intent utterences?

If you would like to contribute multi-intent utterences, please add a + to indicate an additional intent, for example: affirm+ask_transport

Where do I put intents that fall into multiple domains?

Please create a new folder for mixed domains, for example: “mixed-smalltalk+transport”

What about training data that’s not in English?

Currently, we are unable to evaluate the quality of all language contributions, and therefore, during the initial phase we can only accept English training data to the repository. However, we understand that the Rasa community is a global one, and in the long-term we would like to find a solution for this in collaboration with the community.

Why do I need to remove entities from my training data?

We would like to make the training data as easy as possible to adopt to new training models and annotating entities highly dependant on your bot’s purpose. Therefore, we will first focus on collecting training data that only includes intents.

To help you remove the annotated entities from your training data, you can run this script.

About Rasa