DipeshChaudhary/nepali-gector-style-token-level-tag-for-ged

这是一个尼泊尔语的语法错误纠正标注数据集，适用于训练GEC-ToR风格的序列标注模型。数据集包含超过1600万个示例，经过了特殊处理以生成高保真的校正标签，包括复杂的相邻交换操作。数据集分为训练集、验证集和测试集，采用10标签系统进行单词级别的错误标注，并包含了正确句子的标注以提高模型稳定性。

This is a Nepali grammatical error correction annotated dataset designed for training GEC-ToR-style sequence tagging models. The dataset contains over 16 million examples, processed to generate high-fidelity correction tags including complex adjacent swap operations. It is split into training, validation, and test sets, using a 10-tag system for token-level error annotation, and includes correct sentence annotations to enhance model stability.