SQLsynth

该数据集是通过采用人机交互的文本到SQL数据标注系统进行创建的，这种系统显著提升了标注的速度和质量。在评估自然语言问题的正确性和自然度方面，SQLsynth的准确度最高，相较于人工标注和人工智能助手都有更优异的表现。该数据集的任务是进行文本到SQL的数据标注。

This dataset was constructed using a human-machine interactive text-to-SQL data annotation system, which significantly boosts both the speed and quality of the annotation process. In terms of assessing the correctness and naturalness of natural language queries, SQLsynth achieves the highest accuracy, outperforming both manual annotation and AI assistants. The core task of this dataset is text-to-SQL data annotation.