References:
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:xsum')
- Description:
Extreme Summarization (XSum) Dataset.
There are three features:
- document: Input news article.
- summary: One sentence summary of the article.
- id: BBC ID of the article.
- License: No known license
- Version: 1.2.0
- Splits:
Split | Examples |
---|---|
'test' |
11334 |
'train' |
204045 |
'validation' |
11332 |
- Features:
{
"document": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"summary": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"id": {
"dtype": "string",
"id": null,
"_type": "Value"
}
}