rojagtap/natural_questions_clean

---
license: mit
task_categories:
- question-answering
- text-generation
- text2text-generation
language:
- en
tags:
- natural-questions
- question-answering
- text-generation
- text2text
pretty_name: natural-questions-clean
size_categories:
- 100K<n<1M
configs:
- config_name: raw
data_files:
- split: train
path: "raw/train.jsonl"
- split: validation
path: "raw/validation.jsonl"
- config_name: either
data_files:
- split: train
path: "either/train.jsonl"
- split: validation
path: "either/validation.jsonl"
default: true
- config_name: long
data_files:
- split: train
path: "long/train.jsonl"
- split: validation
path: "long/validation.jsonl"
- config_name: short
data_files:
- split: train
path: "short/train.jsonl"
- split: validation
path: "short/validation.jsonl"
---

原始信息汇总

数据集概述

许可证

• MIT许可证

任务类别

• 问答
• 文本生成
• 文本到文本生成

语言

• 英语

标签

• natural-questions
• question-answering
• text-generation
• text2text

数据集名称

• natural-questions-clean

数据集大小

• 100K<n<1M

配置

• config_name: raw

  • 数据文件:
    • 训练集: raw/train.jsonl
    • 验证集: raw/validation.jsonl

• config_name: either

  • 数据文件:
    • 训练集: either/train.jsonl
    • 验证集: either/validation.jsonl
  • 默认配置: 是

• config_name: long

  • 数据文件:
    • 训练集: long/train.jsonl
    • 验证集: long/validation.jsonl

• config_name: short

  • 数据文件:
    • 训练集: short/train.jsonl
    • 验证集: short/validation.jsonl