parallel-sentences-wikimatrix

# Dataset Card for Parallel Sentences - WikiMatrix

This dataset contains parallel sentences (i.e. English sentence + the same sentences in another language) for numerous other languages. Most of the sentences originate from the [OPUS website](https://opus.nlpl.eu/).
In particular, this dataset contains the [WikiMatrix](https://opus.nlpl.eu/WikiMatrix/corpus/version/WikiMatrix) dataset.

## Related Datasets

The following datasets are also a part of the Parallel Sentences collection:
* [parallel-sentences-europarl](https://huggingface.co/datasets/sentence-transformers/parallel-sentences-europarl)
* [parallel-sentences-global-voices](https://huggingface.co/datasets/sentence-transformers/parallel-sentences-global-voices)
* [parallel-sentences-muse](https://huggingface.co/datasets/sentence-transformers/parallel-sentences-muse)
* [parallel-sentences-jw300](https://huggingface.co/datasets/sentence-transformers/parallel-sentences-jw300)
* [parallel-sentences-news-commentary](https://huggingface.co/datasets/sentence-transformers/parallel-sentences-news-commentary)
* [parallel-sentences-opensubtitles](https://huggingface.co/datasets/sentence-transformers/parallel-sentences-opensubtitles)
* [parallel-sentences-talks](https://huggingface.co/datasets/sentence-transformers/parallel-sentences-talks)
* [parallel-sentences-tatoeba](https://huggingface.co/datasets/sentence-transformers/parallel-sentences-tatoeba)
* [parallel-sentences-wikimatrix](https://huggingface.co/datasets/sentence-transformers/parallel-sentences-wikimatrix)
* [parallel-sentences-wikititles](https://huggingface.co/datasets/sentence-transformers/parallel-sentences-wikititles)
* [parallel-sentences-ccmatrix](https://huggingface.co/datasets/sentence-transformers/parallel-sentences-ccmatrix)

These datasets can be used to train multilingual sentence embedding models. For more information, see [sbert.net - Multilingual Models](https://www.sbert.net/examples/training/multilingual/README.html).

## Dataset Subsets

### `all` subset

* Columns: "english", "non_english"
* Column types: `str`, `str`
* Examples:
```python
{
"english": "I will go down with you into Egypt, and I will also surely bring you up again."",
"non_english": "Аз ще бъда с тебе в Египет и Аз ще те изведа назад.""
}
```
* Collection strategy: Combining all other subsets from this dataset.
* Deduplified: No

### `en-...` subsets

* Columns: "english", "non_english"
* Column types: `str`, `str`
* Examples:
```python
{
"english": "By Him Who is the Lord of mankind!",
"non_english": "しかしながら人主の患はまた人を信ぜざるにもある。"
}
```
* Collection strategy: Processing the raw data from [parallel-sentences](https://huggingface.co/datasets/sentence-transformers/parallel-sentences) and formatting it in Parquet, followed by deduplication.
* Deduplified: Yes

# 平行语料数据集卡片 - WikiMatrix

本数据集包含适用于数十种其他语言的平行语料（Parallel Sentences），即英文句子与对应语言的同义句。绝大多数语料源自[OPUS网站](https://opus.nlpl.eu/)。本数据集特别收录了[WikiMatrix](https://opus.nlpl.eu/WikiMatrix/corpus/version/WikiMatrix)数据集。

## 相关数据集
以下数据集同样隶属于平行语料集系列：
* [parallel-sentences-europarl](https://huggingface.co/datasets/sentence-transformers/parallel-sentences-europarl)
* [parallel-sentences-global-voices](https://huggingface.co/datasets/sentence-transformers/parallel-sentences-global-voices)
* [parallel-sentences-muse](https://huggingface.co/datasets/sentence-transformers/parallel-sentences-muse)
* [parallel-sentences-jw300](https://huggingface.co/datasets/sentence-transformers/parallel-sentences-jw300)
* [parallel-sentences-news-commentary](https://huggingface.co/datasets/sentence-transformers/parallel-sentences-news-commentary)
* [parallel-sentences-opensubtitles](https://huggingface.co/datasets/sentence-transformers/parallel-sentences-opensubtitles)
* [parallel-sentences-talks](https://huggingface.co/datasets/sentence-transformers/parallel-sentences-talks)
* [parallel-sentences-tatoeba](https://huggingface.co/datasets/sentence-transformers/parallel-sentences-tatoeba)
* [parallel-sentences-wikimatrix](https://huggingface.co/datasets/sentence-transformers/parallel-sentences-wikimatrix)
* [parallel-sentences-wikititles](https://huggingface.co/datasets/sentence-transformers/parallel-sentences-wikititles)
* [parallel-sentences-ccmatrix](https://huggingface.co/datasets/sentence-transformers/parallel-sentences-ccmatrix)

此类数据集可用于训练多语言句嵌入模型。如需了解更多信息，请参阅[sbert.net - 多语言模型](https://www.sbert.net/examples/training/multilingual/README.html)。

## 数据集子集
### `all` 子集
* 字段：`english`、`non_english`
* 字段类型：字符串、字符串
* 示例：
python
{
"english": "I will go down with you into Egypt, and I will also surely bring you up again.",
"non_english": "Аз ще бъда с тебе в Египет и Аз ще те изведа назад."
}

* 采集策略：合并本数据集的所有其他子集
* 是否去重：否

### `en-...` 子集
* 字段：`english`、`non_english`
* 字段类型：字符串、字符串
* 示例：
python
{
"english": "By Him Who is the Lord of mankind!",
"non_english": "しかしながら人主の患はまた人を信ぜざるにもある。"
}

* 采集策略：先处理[parallel-sentences](https://huggingface.co/datasets/sentence-transformers/parallel-sentences)的原始数据，以Parquet格式进行格式化，随后进行去重
* 是否去重：是