emo

参考：

emo2019

使用以下命令在 TFDS 中加载此数据集：

ds = tfds.load('huggingface:emo/emo2019')
  • 说明：
In this dataset, given a textual dialogue i.e. an utterance along with two previous turns of context, the goal was to infer the underlying emotion of the utterance by choosing from four emotion classes - Happy, Sad, Angry and Others.
  • 许可：无已知许可
  • 版本：1.0.0
  • 拆分：
拆分 样本
'test' 5509
'train' 30160
  • 特征：
{
    "text": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "label": {
        "num_classes": 4,
        "names": [
            "others",
            "happy",
            "sad",
            "angry"
        ],
        "names_file": null,
        "id": null,
        "_type": "ClassLabel"
    }
}