Yugrathee28/hinglish-dataset-1.4M

Hinglish数据集是一个包含1.46百万条Hinglish（印地语和英语混合）评论的工业级NLP数据集，来源于YouTube评论。数据集经过严格的清洗和标注，包含9种意图和10种情感类别，每条评论都有质量评分和标签置信度评分。数据集主要用于研究和评估目的，适用于对话AI、情感分析、低资源NLP研究等场景。The Hinglish Dataset is an industrial-grade NLP dataset containing 1.46 million+ Hinglish (Hindi + English code-mixed) comments sourced from YouTube. The dataset has been rigorously cleaned and labeled, featuring 9 intent classes and 10 emotion classes, with each comment accompanied by a quality score and label confidence score. It is intended strictly for research and evaluation purposes, ideal for conversational AI, sentiment analysis, low-resource NLP research, and more.