Attend the Women in ML Symposium on December 7 Register now

emea

参考:

bg-el

使用以下命令在 TFDS 中加载此数据集:

ds = tfds.load('huggingface:emea/bg-el')
  • 说明
This is a parallel corpus made out of PDF documents from the European Medicines Agency. All files are automatically converted from PDF to plain text using pdftotext with the command line arguments -layout -nopgbrk -eol unix. There are some known problems with tables and multi-column layouts - some of them are fixed in the current version.

source: http://www.emea.europa.eu/

22 languages, 231 bitexts
total number of files: 41,957
total number of tokens: 311.65M
total number of sentence fragments: 26.51M
  • 许可:无已知许可
  • 版本:3.0.0
  • 拆分
拆分 样本
'train' 1044065
  • 特征
{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "languages": [
            "bg",
            "el"
        ],
        "id": null,
        "_type": "Translation"
    }
}

cs-et

使用以下命令在 TFDS 中加载此数据集:

ds = tfds.load('huggingface:emea/cs-et')
  • 说明
This is a parallel corpus made out of PDF documents from the European Medicines Agency. All files are automatically converted from PDF to plain text using pdftotext with the command line arguments -layout -nopgbrk -eol unix. There are some known problems with tables and multi-column layouts - some of them are fixed in the current version.

source: http://www.emea.europa.eu/

22 languages, 231 bitexts
total number of files: 41,957
total number of tokens: 311.65M
total number of sentence fragments: 26.51M
  • 许可:无已知许可
  • 版本:3.0.0
  • 拆分
拆分 样本
'train' 1053164
  • 特征
{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "languages": [
            "cs",
            "et"
        ],
        "id": null,
        "_type": "Translation"
    }
}

de-mt

使用以下命令在 TFDS 中加载此数据集:

ds = tfds.load('huggingface:emea/de-mt')
  • 说明
This is a parallel corpus made out of PDF documents from the European Medicines Agency. All files are automatically converted from PDF to plain text using pdftotext with the command line arguments -layout -nopgbrk -eol unix. There are some known problems with tables and multi-column layouts - some of them are fixed in the current version.

source: http://www.emea.europa.eu/

22 languages, 231 bitexts
total number of files: 41,957
total number of tokens: 311.65M
total number of sentence fragments: 26.51M
  • 许可:无已知许可
  • 版本:3.0.0
  • 拆分
拆分 样本
'train' 1000532
  • 特征
{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "languages": [
            "de",
            "mt"
        ],
        "id": null,
        "_type": "Translation"
    }
}

fr-sk

使用以下命令在 TFDS 中加载此数据集:

ds = tfds.load('huggingface:emea/fr-sk')
  • 说明
This is a parallel corpus made out of PDF documents from the European Medicines Agency. All files are automatically converted from PDF to plain text using pdftotext with the command line arguments -layout -nopgbrk -eol unix. There are some known problems with tables and multi-column layouts - some of them are fixed in the current version.

source: http://www.emea.europa.eu/

22 languages, 231 bitexts
total number of files: 41,957
total number of tokens: 311.65M
total number of sentence fragments: 26.51M
  • 许可:无已知许可
  • 版本:3.0.0
  • 拆分
拆分 样本
'train' 1062753
  • 特征
{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "languages": [
            "fr",
            "sk"
        ],
        "id": null,
        "_type": "Translation"
    }
}

es-lt

使用以下命令在 TFDS 中加载此数据集:

ds = tfds.load('huggingface:emea/es-lt')
  • 说明
This is a parallel corpus made out of PDF documents from the European Medicines Agency. All files are automatically converted from PDF to plain text using pdftotext with the command line arguments -layout -nopgbrk -eol unix. There are some known problems with tables and multi-column layouts - some of them are fixed in the current version.

source: http://www.emea.europa.eu/

22 languages, 231 bitexts
total number of files: 41,957
total number of tokens: 311.65M
total number of sentence fragments: 26.51M
  • 许可:无已知许可
  • 版本:3.0.0
  • 拆分
拆分 样本
'train' 1051370
  • 特征
{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "translation": {
        "languages": [
            "es",
            "lt"
        ],
        "id": null,
        "_type": "Translation"
    }
}