irc_disentangle

参考文献:

ubuntu

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:irc_disentangle/ubuntu')
  • 説明
Disentangling conversations mixed together in a single stream of messages is
a difficult task, made harder by the lack of large manually annotated
datasets. This new dataset of 77,563 messages manually annotated with
reply-structure graphs that both disentangle conversations and define
internal conversation structure. The dataset is 16 times larger than all
previously released datasets combined, the first to include adjudication of
annotation disagreements, and the first to include context.
  • ライセンス: クリエイティブ・コモンズ 表示 4.0 国際公衆利用許諾書
  • バージョン: 1.0.0
  • 分割:
スプリット
'test' 15010
'train' 220616
'validation' 12510
  • 特徴
{
    "id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "raw": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "ascii": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "tokenized": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "date": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "connections": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}

チャンネル2

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:irc_disentangle/channel_two')
  • 説明
Disentangling conversations mixed together in a single stream of messages is
a difficult task, made harder by the lack of large manually annotated
datasets. This new dataset of 77,563 messages manually annotated with
reply-structure graphs that both disentangle conversations and define
internal conversation structure. The dataset is 16 times larger than all
previously released datasets combined, the first to include adjudication of
annotation disagreements, and the first to include context.
  • ライセンス: クリエイティブ・コモンズ 表示 4.0 国際公衆利用許諾書
  • バージョン: 1.0.0
  • 分割:
スプリット
'all_' 2602
'dev' 1001
'pilot' 501
'pilot_dev' 1501
'test' 1001
  • 特徴
{
    "id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    },
    "raw": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "ascii": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "tokenized": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "connections": {
        "feature": {
            "dtype": "int32",
            "id": null,
            "_type": "Value"
        },
        "length": -1,
        "id": null,
        "_type": "Sequence"
    }
}