Question 1

What is the Conversational Text-to-SQL Systems (CoSQL) dataset?

Accepted Answer

Dataset consists of 30k+ turns plus 10k+ annotated SQL queries, obtained from a Wizard-of-Oz collection of 3k dialogues querying 200 complex databases spanning 138 domains.It is the dilaogue version of the Spider and SParC tasks.

Question 2

Is Conversational Text-to-SQL Systems (CoSQL) a benchmark?

Accepted Answer

Conversational Text-to-SQL Systems (CoSQL) is a dataset for training or evaluation; it isn't tracked as a standard LLM benchmark in our catalog.

Question 3

Where can I download Conversational Text-to-SQL Systems (CoSQL)?

Accepted Answer

Conversational Text-to-SQL Systems (CoSQL) is available at its source: https://yale-lily.github.io/cosql.

Conversational Text-to-SQL Systems (CoSQL)

About Conversational Text-to-SQL Systems (CoSQL)

Details

FAQ