Thanmay/belebele

该数据集包含多个配置，如默认配置、英语拉丁字母配置和印地语天城文配置。每个配置都有相同的特征，包括链接、问题编号、段落、问题、多项选择题答案、正确答案编号、方言、时间戳、标题和最近页面等。数据集主要用于训练，每个配置的训练集大小和示例数量也有所不同。

The dataset contains multiple configurations such as default, English Latin script, and Hindi Devanagari script. Each configuration has the same features, including link, question number, passage, question, multiple-choice answers, correct answer number, dialect, timestamp, title, and nearest pages. The dataset is primarily used for training, with varying sizes and numbers of examples for each configurations training set.

原始信息汇总

数据集概述

数据集配置

默认配置 (default)

• 特征:
  • link: string
  • question_number: int64
  • flores_passage: string
  • question: string
  • mc_answer1: string
  • mc_answer2: string
  • mc_answer3: string
  • mc_answer4: string
  • correct_answer_num: string
  • dialect: string
  • ds: timestamp[s]
  • title: string
  • nearest_pages: sequence[string]
• 分割:
  • train:
    • num_bytes: 23467133
    • num_examples: 900
• 下载大小: 9496682
• 数据集大小: 23467133
• 数据文件路径: data/train-*

英语配置 (eng_Latn)

• 特征:
  • link: string
  • question_number: int64
  • flores_passage: string
  • question: string
  • mc_answer1: string
  • mc_answer2: string
  • mc_answer3: string
  • mc_answer4: string
  • correct_answer_num: string
  • dialect: string
  • ds: timestamp[s]
  • title: string
• 分割:
  • train:
    • num_bytes: 747147
    • num_examples: 900
• 下载大小: 312963
• 数据集大小: 747147
• 数据文件路径: eng_Latn/train-*

印地语配置 (hin_Deva)

• 特征:
  • link: string
  • question_number: int64
  • flores_passage: string
  • question: string
  • mc_answer1: string
  • mc_answer2: string
  • mc_answer3: string
  • mc_answer4: string
  • correct_answer_num: string
  • dialect: string
  • ds: timestamp[s]
  • title: string
• 分割:
  • train:
    • num_bytes: 747147
    • num_examples: 900
• 下载大小: 315961
• 数据集大小: 747147
• 数据文件路径: hin_Deva/train-*