Some questions about datasets

[yefenghua]

How do these files work in the project? (it is a bit different from your essay)

./PROGRAM_DEMO/data/

1. data
2. dummy_data
    2.1.dstc
    2.2.msmarco
    2.3.reddit
    2.4.unifedqa
3.grounded
4.ungrounded

Also, I wonder how can I train the project with my own datasets?

I wonder if I am right to use the EXAMPLE.JSON to train directly. If it is wrong, how can I do it CORRECTLY?

  {
    "Context": "Please remind me of calling to Jessie at 2PM.",
    "Knowledge": "reminder_contact_name is Jessie, reminder_time is 2PM",
    "Response": "Sure, set the reminder: call to Jesse at 2PM"
  },

[ba2512005]

@yefenghua You might want to look into a Text-to-intent model like https://github.com/thuiar/TEXTOIR