Referencias:
en-hr
Utilice el siguiente comando para cargar este conjunto de datos en TFDS:
ds = tfds.load('huggingface:opus_tedtalks/en-hr')
- Descripción :
This is a Croatian-English parallel corpus of transcribed and translated TED talks, originally extracted from https://wit3.fbk.eu. The corpus is compiled by Željko Agić and is taken from http://lt.ffzg.hr/zagic provided under the CC-BY-NC-SA license.
2 languages, total number of files: 2
total number of tokens: 2.81M
total number of sentence fragments: 0.17M
- Licencia : Ninguna licencia conocida
- Versión : 1.0.0
- Divisiones :
Dividir | Ejemplos |
---|---|
'train' | 86348 |
- Características :
{
"id": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"translation": {
"languages": [
"en",
"hr"
],
"id": null,
"_type": "Translation"
}
}