---
dataset_info:
  features:
  - name: scenes
    sequence:
      sequence: float32
      length: 28
    length: 35
  - name: language_annotation
    dtype: string
  - name: subdataset_idx
    dtype: int64
  splits:
  - name: train
    num_bytes: 130702040623
    num_examples: 32523270
  download_size: 10341801176
  dataset_size: 130702040623
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---