ro_sts

参考文献:

ro_sts

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:ro_sts/ro_sts')
  • 説明
The RO-STS (Romanian Semantic Textual Similarity) dataset contains 8628 pairs of sentences with their similarity score. It is a high-quality translation of the STS benchmark dataset.
  • ライセンス: CC BY-SA 4.0 ライセンス
  • バージョン: 1.0.0
  • 分割:
スプリット
'test' 1379年
'train' 5749
'validation' 1500
  • 特徴
{
    "score": {
        "dtype": "float32",
        "id": null,
        "_type": "Value"
    },
    "sentence1": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence2": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}