---
dataset_info:
  features:
  - name: img
    dtype: image
  - name: human_score
    sequence: float32
  - name: cand
    dtype: string
  splits:
  - name: train
    num_bytes: 41002482155.375
    num_examples: 81909
  - name: validation
    num_bytes: 14600600877.5
    num_examples: 28852
  - name: test
    num_bytes: 14479234747.75
    num_examples: 28842
  download_size: 69347999824
  dataset_size: 70082317780.625
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
  - split: test
    path: data/test-*
---