---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: safety_label
    dtype: string
  - name: category
    dtype: string
  - name: rationale
    dtype: string
  - name: source
    dtype: string
  - name: data_source
    dtype: string
  splits:
  - name: train
    num_bytes: 2401583719.038
    num_examples: 22161
  download_size: 1551110405
  dataset_size: 2401583719.038
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---