---
dataset_info:
  features:
  - name: scenes
    sequence:
      sequence: float32
  - name: language_annotation
    dtype: string
  splits:
  - name: train
    num_bytes: 128534154742
    num_examples: 18515820
  download_size: 16173858076
  dataset_size: 128534154742
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---