README.md

metadata

license: cc-by-sa-4.0
dataset_info:
  features:
    - name: premise
      dtype: string
    - name: hypothesis
      dtype: string
    - name: label
      dtype: string
    - name: source
      dtype: string
    - name: org_premise
      dtype: string
    - name: org_hypothesis
      dtype: string
    - name: __index_level_0__
      dtype: int64
  splits:
    - name: train
      num_bytes: 12554097
      num_examples: 40000
    - name: dev
      num_bytes: 1569723
      num_examples: 4961
    - name: test
      num_bytes: 1584740
      num_examples: 5000
  download_size: 8471333
  dataset_size: 15708560

Slovene MNLI SNLI

This dataset contains 49961 premise hypothesis pairs (50% MNLI, 50% SNLI), which were acquired by translating original samples.