参考：
curiosity_dialogs
使用以下命令在 TFDS 中加载此数据集：
ds = tfds.load('huggingface:curiosity_dialogs/curiosity_dialogs')
- 说明：
This dataset contains 14K dialogs (181K utterances) where users and assistants converse about geographic topics like
geopolitical entities and locations. This dataset is annotated with pre-existing user knowledge, message-level dialog
acts, grounding to Wikipedia, and user reactions to messages.
| 拆分 | 样本 |
|---|---|
'test' |
1287 |
'test_zero' |
1187 |
'train' |
10287 |
'val' |
1287 |
- 特征：
{
"messages": {
"feature": {
"message": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"liked": {
"num_classes": 2,
"names": [
"False",
"True"
],
"names_file": null,
"id": null,
"_type": "ClassLabel"
},
"sender": {
"num_classes": 2,
"names": [
"user",
"assistant"
],
"names_file": null,
"id": null,
"_type": "ClassLabel"
},
"facts": {
"feature": {
"fid": {
"dtype": "int32",
"id": null,
"_type": "Value"
},
"used": {
"num_classes": 2,
"names": [
"False",
"True"
],
"names_file": null,
"id": null,
"_type": "ClassLabel"
},
"source": {
"num_classes": 3,
"names": [
"section",
"known",
"random"
],
"names_file": null,
"id": null,
"_type": "ClassLabel"
}
},
"length": -1,
"id": null,
"_type": "Sequence"
},
"message_id": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"dialog_acts": {
"feature": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"length": -1,
"id": null,
"_type": "Sequence"
}
},
"length": -1,
"id": null,
"_type": "Sequence"
},
"known_entities": {
"feature": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"length": -1,
"id": null,
"_type": "Sequence"
},
"focus_entity": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"dialog_id": {
"dtype": "int32",
"id": null,
"_type": "Value"
},
"inferred_steps": {
"num_classes": 2,
"names": [
"False",
"True"
],
"names_file": null,
"id": null,
"_type": "ClassLabel"
},
"created_time": {
"dtype": "int64",
"id": null,
"_type": "Value"
},
"aspects": {
"feature": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"length": -1,
"id": null,
"_type": "Sequence"
},
"first_aspect": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"second_aspect": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"shuffle_facts": {
"num_classes": 2,
"names": [
"False",
"True"
],
"names_file": null,
"id": null,
"_type": "ClassLabel"
},
"related_entities": {
"feature": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"length": -1,
"id": null,
"_type": "Sequence"
},
"tag": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"user_id": {
"dtype": "int32",
"id": null,
"_type": "Value"
},
"assistant_id": {
"dtype": "int32",
"id": null,
"_type": "Value"
},
"is_annotated": {
"num_classes": 2,
"names": [
"False",
"True"
],
"names_file": null,
"id": null,
"_type": "ClassLabel"
},
"user_dialog_rating": {
"dtype": "int32",
"id": null,
"_type": "Value"
},
"user_other_agent_rating": {
"dtype": "int32",
"id": null,
"_type": "Value"
},
"assistant_dialog_rating": {
"dtype": "int32",
"id": null,
"_type": "Value"
},
"assistant_other_agent_rating": {
"dtype": "int32",
"id": null,
"_type": "Value"
},
"reported": {
"num_classes": 2,
"names": [
"False",
"True"
],
"names_file": null,
"id": null,
"_type": "ClassLabel"
},
"annotated": {
"num_classes": 2,
"names": [
"False",
"True"
],
"names_file": null,
"id": null,
"_type": "ClassLabel"
}
}