lm1b

参考文献:

平文

次のコマンドを使用して、このデータセットを TFDS にロードします。

ds = tfds.load('huggingface:lm1b/plain_text')
  • 説明
A benchmark corpus to be used for measuring progress in statistical language modeling. This has almost one billion words in the training data.
  • ライセンス: 不明なライセンス
  • バージョン: 1.0.0
  • 分割:
スプリット
'test' 306688
'train' 30301028
  • 特徴
{
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}