cdleong/piglatin-mt

---
language:
- en
license:
- mit
multilinguality:
- translation
size_categories:
- 10K<n<100K
source_datasets:
- original
task_categories:
- translation
task_ids: []
language_details: eng and engyay
---
## Dataset Description

- **Homepage:** cdleong.github.io

# Dataset Summary:
Pig-latin machine and English parallel machine translation corpus.

Based on [The Project Gutenberg EBook of "De Bello Gallico" and Other Commentaries](https://www.gutenberg.org/ebooks/10657)

Converted to pig-latin with https://github.com/bpabel/piglatin

Blank lines removed.

## Dataset Structure

```
DatasetDict({
train: Dataset({
features: ['translation'],
num_rows: 14778
})
validation: Dataset({
features: ['translation'],
num_rows: 1000
})
})
```

### Data Instances
```
{
'translation':
{
'eng': 'thrown into disorder they returned with more precipitation than is usual',
'engyay': 'own-thray into-ay isorder-day ey-thay eturned-ray ith-way ore-may ecipitation-pray an-thay is-ay usual-ay'
}
}
```

### Data Fields

- `translation`: a dictionary containing two strings paired with a key indicating the corresponding language.

### Data Splits
- `train`: most of the data, 13,232 samples total.
- `dev`: 1k holdout samples, created with the datasets.train_test_split() function

原始信息汇总

数据集概述

基本信息

• 语言: 英语（eng）和英语拉丁语（engyay）
• 许可证: MIT
• 多语言性: 翻译
• 大小: 10K<n<100K
• 源数据集: 原始数据
• 任务类别: 翻译

数据集描述

• 摘要: 猪拉丁语与英语平行机器翻译语料库。
• 来源: 基于《The Project Gutenberg EBook of "De Bello Gallico" and Other Commentaries》，转换为猪拉丁语。

数据集结构

• 数据集字典: 包含训练集和验证集
  • 训练集: 包含14,778行数据，特征为translation
  • 验证集: 包含1,000行数据，特征为translation

数据实例

• 示例:

  { translation: { eng: thrown into disorder they returned with more precipitation than is usual, engyay: own-thray into-ay isorder-day ey-thay eturned-ray ith-way ore-may ecipitation-pray an-thay is-ay usual-ay } }

数据字段

• translation: 包含两个字符串，分别对应英语和英语拉丁语。

数据分割

• 训练集: 13,232个样本
• 验证集: 1,000个样本，通过datasets.train_test_split()函数创建