---
dataset_info:
  features:
  - name: gen_image
    dtype: image
  - name: ref_image
    dtype: image
  - name: sscd_score
    dtype: float32
  - name: caption
    dtype: string
  - name: original_url
    dtype: string
  - name: index
    dtype: int32
  splits:
  - name: train
    num_bytes: 396090953.0
    num_examples: 454
  download_size: 394985954
  dataset_size: 396090953.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---