kilicai/turkish-sft-clean-v4-final

这是一个土耳其语监督微调（SFT）数据集，用于文本生成任务，包含10万到100万条数据。该版本为v4最终版，但已被取代，因为它存在一些内容级别的问题，如类别标签与任务不匹配、重复内容等。建议使用更严格的v5版本以获得更高质量的数据。

This is a Turkish supervised fine-tuning (SFT) dataset for text generation tasks, containing between 100,000 and 1,000,000 examples. This version is v4 final, but it has been superseded due to content-level issues such as mismatched category labels and duplicate content. It is recommended to use the stricter v5 version for higher quality data.