PLUS: Opensource text to SQL code model outperforms GPT-4, Test multiple VLMs side-by-side
>The development of SQLCoder involved a curated dataset comprising more than 20,000 questions based on 10 different schemas
I couldn't find any reference to 20K questions. Is it private?
>The development of SQLCoder involved a curated dataset comprising more than 20,000 questions based on 10 different schemas
I couldn't find any reference to 20K questions. Is it private?