---
dataset_info:
  features:
  - name: img
    dtype: image
  - name: image_id
    dtype: int64
  - name: caption
    dtype: string
  - name: identity_id
    dtype: int64
  - name: similarity
    dtype: float32
  - name: variant
    dtype:
      class_label:
        names:
          '0': headshot
          '1': fullbody
          '2': side
          '3': midshot
          '4': behind
  - name: original_filename
    dtype: string
  splits:
  - name: train
    num_bytes: 172029924
    num_examples: 450
  - name: test
    num_bytes: 18997819
    num_examples: 50
  download_size: 190983158
  dataset_size: 191027743
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
---