---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: flickr_id
    dtype: string
  - name: hypothesis
    dtype: string
  - name: gold_label
    dtype:
      class_label:
        names:
          '0': entailment
          '1': neutral
          '2': contradiction
  splits:
  - name: test
    num_bytes: 2021696174.84
    num_examples: 14740
  - name: dev
    num_bytes: 1996850972.232
    num_examples: 14339
  - name: train
    num_bytes: 57824175716.576
    num_examples: 401717
  download_size: 8309256949
  dataset_size: 61842722863.647995
configs:
- config_name: default
  data_files:
  - split: test
    path: data/test-*
  - split: dev
    path: data/dev-*
  - split: train
    path: data/train-*
---