wmt19_translate

  • Description:

Translate dataset based on the data from statmt.org.

Versions exists for the different years using a combination of multiple data sources. The base wmt_translate allows you to create your own config to choose your own data/language pair by creating a custom tfds.translate.wmt.WmtConfig.

config = tfds.translate.wmt.WmtConfig(
    version="0.0.1",
    language_pair=("fr", "de"),
    subsets={
        tfds.Split.TRAIN: ["commoncrawl_frde"],
        tfds.Split.VALIDATION: ["euelections_dev2019"],
    },
)
builder = tfds.builder("wmt_translate", config=config)
  • Homepage: http://www.statmt.org/wmt19/translation-task.html
  • Source code: tfds.translate.wmt19.Wmt19Translate
  • Versions:
    • 1.0.0 (default): No release notes.
  • Dataset size: Unknown size
  • Manual download instructions: This dataset requires you to download the source data manually into download_config.manual_dir (defaults to ~/tensorflow_datasets/manual/):
    Some of the wmt configs here, require a manual download. Please look into wmt.py to see the exact path (and file name) that has to be downloaded.
  • Auto-cached (documentation): No
  • Citation:
@ONLINE {wmt19translate,
    author = "Wikimedia Foundation",
    title  = "ACL 2019 Fourth Conference on Machine Translation (WMT19), Shared Task: Machine Translation of News",
    url    = "http://www.statmt.org/wmt19/translation-task.html"
}

wmt19_translate/cs-en (default config)

  • Config description: WMT 2019 cs-en translation task dataset.

  • Download size: 1.88 GiB

  • Splits:

Split Examples
'train' 20,246,548
'validation' 2,983
  • Features:
Translation({
    'cs': Text(shape=(), dtype=tf.string),
    'en': Text(shape=(), dtype=tf.string),
})

wmt19_translate/de-en

  • Config description: WMT 2019 de-en translation task dataset.

  • Download size: 9.71 GiB

  • Splits:

Split Examples
'train' 38,690,334
'validation' 2,998
  • Features:
Translation({
    'de': Text(shape=(), dtype=tf.string),
    'en': Text(shape=(), dtype=tf.string),
})

wmt19_translate/fi-en

  • Config description: WMT 2019 fi-en translation task dataset.

  • Download size: 959.46 MiB

  • Splits:

Split Examples
'train' 6,587,448
'validation' 3,000
  • Features:
Translation({
    'en': Text(shape=(), dtype=tf.string),
    'fi': Text(shape=(), dtype=tf.string),
})

wmt19_translate/gu-en

  • Config description: WMT 2019 gu-en translation task dataset.

  • Download size: 37.03 MiB

  • Splits:

Split Examples
'train' 11,670
'validation' 1,998
  • Features:
Translation({
    'en': Text(shape=(), dtype=tf.string),
    'gu': Text(shape=(), dtype=tf.string),
})

wmt19_translate/kk-en

  • Config description: WMT 2019 kk-en translation task dataset.

  • Download size: 39.58 MiB

  • Splits:

Split Examples
'train' 126,583
'validation' 2,066
  • Features:
Translation({
    'en': Text(shape=(), dtype=tf.string),
    'kk': Text(shape=(), dtype=tf.string),
})

wmt19_translate/lt-en

  • Config description: WMT 2019 lt-en translation task dataset.

  • Download size: 392.20 MiB

  • Splits:

Split Examples
'train' 2,344,893
'validation' 2,000
  • Features:
Translation({
    'en': Text(shape=(), dtype=tf.string),
    'lt': Text(shape=(), dtype=tf.string),
})

wmt19_translate/ru-en

  • Config description: WMT 2019 ru-en translation task dataset.

  • Download size: 3.86 GiB

  • Splits:

Split Examples
'train' 38,492,126
'validation' 3,000
  • Features:
Translation({
    'en': Text(shape=(), dtype=tf.string),
    'ru': Text(shape=(), dtype=tf.string),
})

wmt19_translate/zh-en

  • Config description: WMT 2019 zh-en translation task dataset.

  • Download size: 2.04 GiB

  • Splits:

Split Examples
'train' 25,986,436
'validation' 3,981
  • Features:
Translation({
    'en': Text(shape=(), dtype=tf.string),
    'zh': Text(shape=(), dtype=tf.string),
})

wmt19_translate/fr-de

  • Config description: WMT 2019 fr-de translation task dataset.

  • Download size: 722.20 MiB

  • Splits:

Split Examples
'train' 9,824,476
'validation' 1,512
  • Features:
Translation({
    'de': Text(shape=(), dtype=tf.string),
    'fr': Text(shape=(), dtype=tf.string),
})