snli, snli

  • Descrição :

O corpus SNLI (versão 1.0) é uma coleção de 570.000 pares de sentenças em inglês escritas por humanos rotuladas manualmente para classificação balanceada com os rótulos implicação, contradição e neutro, suportando a tarefa de inferência de linguagem natural (NLI), também conhecida como reconhecimento de implicação textual (RTE).

Dividir Exemplos
'test' 10.000
'train' 550.152
'validation' 10.000
  • Estrutura de recursos :
FeaturesDict({
    'hypothesis': Text(shape=(), dtype=string),
    'label': ClassLabel(shape=(), dtype=int64, num_classes=3),
    'premise': Text(shape=(), dtype=string),
})
  • Documentação do recurso:
Característica Classe Forma Tipo D Descrição
RecursosDict
hipótese Texto corda
rótulo ClassLabel int64
premissa Texto corda
  • Citação :
@inproceedings{snli:emnlp2015,
    Author = {Bowman, Samuel R. and Angeli, Gabor and Potts, Christopher, and Manning, Christopher D.},
    Booktitle = {Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing (EMNLP)},
    Publisher = {Association for Computational Linguistics},
    Title = {A large annotated corpus for learning natural language inference},
    Year = {2015}
}