Referencias:
Utilice el siguiente comando para cargar este conjunto de datos en TFDS:
ds = tfds.load('huggingface:limit')
- Descripción :
Motion recognition is one of the basic cognitive capabilities of many life forms, yet identifying motion of physical entities in natural language have not been explored extensively and empirically. Literal-Motion-in-Text (LiMiT) dataset, is a large human-annotated collection of English text sentences describing physical occurrence of motion, with annotated physical entities in motion.
- Licencia : Sin licencia conocida
- Versión : 1.0.0
- Divisiones :
Separar | Ejemplos |
---|---|
'test' | 1000 |
'train' | 23559 |
- Características :
{
"id": {
"dtype": "int32",
"id": null,
"_type": "Value"
},
"sentence": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"motion": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"motion_entities": [
{
"entity": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"start_index": {
"dtype": "int32",
"id": null,
"_type": "Value"
}
}
]
}