---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: mask
    dtype: image
  - name: masked_image
    dtype: image
  - name: mae_embedding
    dtype:
      array3_d:
        shape:
        - 1
        - 196
        - 768
        dtype: float32
  splits:
  - name: train
    num_bytes: 13629084978.0
    num_examples: 1000
  download_size: 13640186121
  dataset_size: 13629084978.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---