---
dataset_info:
  features:
  - name: scenes
    sequence:
      sequence: float32
  - name: language_annotation
    dtype: string
  splits:
  - name: train
    num_bytes: 41820172461
    num_examples: 7966221
  download_size: 4545543118
  dataset_size: 41820172461
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---