---
dataset_info:
  features:
  - name: scenes
    sequence:
      sequence: float32
  - name: language_annotation
    dtype: string
  splits:
  - name: train
    num_bytes: 82817658454
    num_examples: 27262539
  download_size: 5007333149
  dataset_size: 82817658454
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---