bijode8658-kudimi/sft-2

Sft 2数据集包含多个配置，如chunked、ingested、multi_hop_questions、prepared_lighteval、single_shot_questions和summarized。每个配置具有不同的特征，旨在支持不同的自然语言处理任务，如文档摘要、问题生成和多跳问题回答。数据集包含文档ID、文本、文件名、元数据、摘要、问题、答案等多种信息，并提供了训练集的分割和大小详情。The Sft 2 dataset includes multiple configurations such as chunked, ingested, multi_hop_questions, prepared_lighteval, single_shot_questions, and summarized. Each configuration has distinct features designed to support various NLP tasks like document summarization, question generation, and multi-hop question answering. The dataset contains various types of information including document IDs, text, filenames, metadata, summaries, questions, and answers, along with details on train splits and sizes.