请问可以开源spider2.0的评估数据吗？

[xingyuma618]

您好，关注到您的工作非常感兴趣，想测试您这边的spider2.0的效果，看到您这边是有对spider2.0单独处理了sqlite的数据，请问可以开源处理好的数据或者处理脚本吗？

[xingyuma618]

例如评估代码里用到的这几个数据，是spider2.0官方数据里没有的，可能是您这边处理好的数据？期待您的回复！

[lihaoyang-ruc]

You can download the datasets used in OmniSQL from any of the following sources:
https://modelscope.cn/datasets/seeklhy/OmniSQL-datasets/summary
https://huggingface.co/datasets/seeklhy/OmniSQL-datasets

The datasets include BIRD, Spider, ScienceBenchmark, EHRSQL, Spider2-SQLite, Spider-DK, Spider-Realistic, Spider-Syn, and SynSQL-2.5M.

These instructions have been provided in the README file. Please refer to:
https://github.com/RUCKBReasoning/OmniSQL/tree/main/train_and_evaluate#dataset-preparation