Skip to content

ToTTo

Table-to-TextEnglish

ToTTo is a table-to-text-focused dataset in English that provides 120,000+ labeled examples distributed in JSON format.

About ToTTo

Dataset is used for the controlled generation of descriptions of tabular data comprising over 100,000 examples. Each example is a aligned pair of a highlighted table and the description of the highlighted content.

Details

Task
Table-to-Text
Language
English
Format
JSON
Rows / instances
120,000+
Creator
Parikh et al.
Year
2020
Download Paper

FAQ