中文版本的 README

Natural Language Understanding Benchmarks Datasets

This project has collected some public data sets that can be used for NLU benchmark-testing.

data set

DialogFlow

Summary

intents

Count: 11

'weather.activity'
'weather.temperature'
'weather - context:weather - comment:outfit'
'weather'
'weather - context:weather - comment:address & date-time'
'weather.outfit'
'weather - context:weather - comment:address'
'weather - context:weather - comment:date time'
'weather - context:weather - comment:condition'
'weather - context:weather - comment:activity'
'weather.condition'

entities

Count: 7

'activity'
'condition'
'outfit'
'address'
'temperature'
'unit'
'date-time'

Download

Language	Dataset	DialogFlow Format	Rasa NLU Format
Chinese	small-talk	zh-CN.zip	zh-CN.zip
Chinese	support	zh-CN.zip	zh-CN.zip
Chinese	translate	zh-CN.zip	zh-CN.zip
Chinese	weather	zh-CN.zip	zh-CN.zip
English	small-talk	en.zip	en.zip
English	support	en.zip	en.zip
English	Translate	en.zip	en.zip
English	weather	en.zip	en.zip

* Copyright DialogFlow

ATIS Data

TODO