Gabriel/gigaword_swe

---
language:
- sv
license:
- mit
size_categories:
- 1M<n<3M
source_datasets:
- https://github.com/huggingface/datasets/tree/master/datasets/gigaword
task_categories:
- summarization
- text2text-generation
task_ids: []
tags:
- conditional-text-generation
---

# Dataset Card for Swedish Gigaword Dataset
The Swedish gigaword dataset has only been machine-translated to improve downstream fine-tuning on Swedish summarization tasks.

## Dataset Summary
Read about the full details at original English version: https://huggingface.co/datasets/gigaword

### Data Fields
- `document`: a string containing the shorter body
- `summary`: a string containing the summary of the body

### Data Splits
The Swedish gigaword dataset follows the same splits as the original English version and has 3 splits: _train_, _validation_, and _test_.

| Dataset Split | Number of Instances in Split |
| ------------- | ------------------------------------------- |
| Train | 3,700,301 |
| Validation | 189,650 |
| Test | 1,951 |

原始信息汇总

瑞典语Gigaword数据集概述

数据集基本信息

• 语言: 瑞典语 (sv)
• 许可证: MIT
• 大小: 1M<n<3M
• 源数据集: Gigaword
• 任务类别: 摘要生成, 文本到文本生成
• 标签: 条件文本生成

数据集描述

瑞典语Gigaword数据集主要用于改进瑞典语摘要任务的下游微调。

数据字段

• document: 包含较短正文的字符串
• summary: 包含正文摘要的字符串

数据分割

数据集遵循与原始英语版本相同的分割方式，包含三个分割：训练集、验证集和测试集。

数据集分割	分割中的实例数量
训练集	3,700,301
验证集	189,650
测试集	1,951