NSMC

参考文献:

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:nsmc')
  • 説明
This is a movie review dataset in the Korean language. Reviews were scraped from Naver movies. The dataset construction is based on the method noted in Large movie review dataset from Maas et al., 2011.
  • ライセンス: CC0 1.0 ユニバーサル (CC0 1.0)
  • バージョン: 1.1.0
  • 分割:
スプリット
'test' 50000
'train' 150000
  • 特徴
{
    "id": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "document": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "label": {
        "num_classes": 2,
        "names": [
            "negative",
            "positive"
        ],
        "names_file": null,
        "id": null,
        "_type": "ClassLabel"
    }
}