---
dataset_info:
  features:
  - name: query_image
    dtype: image
  - name: labels
    list: int64
  - name: concept_images
    list: image
  - name: num_concepts
    dtype: string
  - name: dialogues_json
    dtype: string
  - name: qas_json
    dtype: string
  - name: text
    dtype: string
  splits:
  - name: train
    num_bytes: 4633455605
    num_examples: 2796
  - name: test
    num_bytes: 2451239274
    num_examples: 1386
  download_size: 7247441170
  dataset_size: 7084694879
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
---