Referencias:
ro_sts
Utilice el siguiente comando para cargar este conjunto de datos en TFDS:
ds = tfds.load('huggingface:ro_sts/ro_sts')
- Descripción :
The RO-STS (Romanian Semantic Textual Similarity) dataset contains 8628 pairs of sentences with their similarity score. It is a high-quality translation of the STS benchmark dataset.
- Licencia : Licencia CC BY-SA 4.0
- Versión : 1.0.0
- Divisiones :
| Separar | Ejemplos |
|---|---|
'test' | 1379 |
'train' | 5749 |
'validation' | 1500 |
- Características :
{
"score": {
"dtype": "float32",
"id": null,
"_type": "Value"
},
"sentence1": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"sentence2": {
"dtype": "string",
"id": null,
"_type": "Value"
}
}