decodingchris/clean_squad_classic_v1
收藏Hugging Face2025-01-30 更新2025-02-15 收录
下载链接：
https://hf-mirror.com/datasets/decodingchris/clean_squad_classic_v1
下载链接
链接失效反馈官方服务：
资源简介：
Clean SQuAD Classic v1数据集是原始SQuAD v1数据集的精炼版，经过预处理以提高自然语言处理任务（如问题回答）的数据质量和可用性。该数据集通过去除问题字段的首尾空格和过滤掉过短的问题进行了优化。数据集包含训练和验证两个子集，每个子集都包括问题-上下文对的唯一标识符、文章标题、上下文、预处理后的问题以及正确答案的文本和答案在上下文中的起始位置。
The Clean SQuAD Classic v1 dataset is a refined version of the original SQuAD v1 dataset, preprocessed to enhance data quality and usability for NLP tasks such as Question Answering. The dataset has been optimized by trimming whitespace from the question field and filtering out questions that are shorter than 12 characters. It includes two subsets, train and validation, each containing unique identifiers for the question-context pairs, article titles, contexts, preprocessed questions, and the text of the correct answers along with the starting position of the answer in the context.
提供机构：
decodingchris
