large-traversaal/urdu-instruct
收藏Hugging Face2025-10-13 更新2025-10-18 收录
下载链接：
https://hf-mirror.com/datasets/large-traversaal/urdu-instruct
下载链接
链接失效反馈官方服务：
资源简介：
Urdu-Instruct数据集是一个包含51,686个训练示例和1,084个测试示例的高质量多语种合成语料库，旨在提高乌尔都语中的指令遵循、推理和双语理解能力。该数据集支持七种乌尔都语言任务：生成、推理、伦理、问答、翻译、分类和情感分析。
The Urdu-Instruct dataset is a high-quality multilingual synthetic corpus containing 51,686 training and 1,084 test examples, designed to improve instruction-following, reasoning, and bilingual understanding in Urdu. It supports seven Urdu language tasks: Generation, Reasoning, Ethics, Question Answering, Translation, Classification, and Sentiment Analysis.
提供机构：
large-traversaal
