README.md

metadata

dataset_info:
  features:
    - name: premise
      dtype: large_string
    - name: hypothesis
      dtype: large_string
    - name: label
      dtype:
        class_label:
          names:
            '0': entailment
            '1': neutral
            '2': contradiction
  splits:
    - name: train
      num_bytes: 2026433
      num_examples: 12667
    - name: validation
      num_bytes: 257194
      num_examples: 1583
    - name: test
      num_bytes: 250935
      num_examples: 1584
  download_size: 1150302
  dataset_size: 2534562
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*
      - split: test
        path: data/test-*