bbc-hindi-nli

Dataset for Natural Language Inference in Hindi Language. BBC Hindi Dataset consists of textual-entailment pairs. Each row of the Datasets if made up of 4 columns - Premise, Hypothesis, Label and Topic. Context and Hypothesis is written in Hindi while Entailment-Label is in English. Entailment-label is of 2 types - entailed and not-entailed. Dataset can be used to train models for Natural Language Inference tasks in Hindi Language.