参考：
dialog_re
使用以下命令在 TFDS 中加载此数据集：
ds = tfds.load('huggingface:dialog_re/dialog_re')
- 说明：
DialogRE is the first human-annotated dialogue based relation extraction (RE) dataset aiming
to support the prediction of relation(s) between two arguments that appear in a dialogue.
The dataset annotates all occurrences of 36 possible relation types that exist between pairs
of arguments in the 1,788 dialogues originating from the complete transcripts of Friends.
- 许可：https://github.com/nlpdata/dialogre/blob/master/license.txt
- 版本：1.0.0
- 拆分：
| 拆分 | 样本 |
|---|---|
'test' |
357 |
'train' |
1073 |
'validation' |
358 |
- 特征：
{
"dialog": {
"feature": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"length": -1,
"id": null,
"_type": "Sequence"
},
"relation_data": {
"feature": {
"x": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"y": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"x_type": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"y_type": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"r": {
"feature": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"length": -1,
"id": null,
"_type": "Sequence"
},
"rid": {
"feature": {
"dtype": "int32",
"id": null,
"_type": "Value"
},
"length": -1,
"id": null,
"_type": "Sequence"
},
"t": {
"feature": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"length": -1,
"id": null,
"_type": "Sequence"
}
},
"length": -1,
"id": null,
"_type": "Sequence"
}
}