---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: original_index
    dtype: int64
  - name: landmark
    sequence:
      sequence: int64
  - name: mask
    dtype: 'null'
  splits:
  - name: train
    num_bytes: 3708542193.5
    num_examples: 52812
  download_size: 3632739568
  dataset_size: 3708542193.5
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---