README.md

---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: landmarks
    struct:
    - name: nose
      struct:
      - name: x
        dtype: float32
      - name: y
        dtype: float32
    - name: r_eye
      struct:
      - name: x
        dtype: float32
      - name: y
        dtype: float32
    - name: l_eye
      struct:
      - name: x
        dtype: float32
      - name: y
        dtype: float32
    - name: r_ear
      struct:
      - name: x
        dtype: float32
      - name: y
        dtype: float32
    - name: l_ear
      struct:
      - name: x
        dtype: float32
      - name: y
        dtype: float32
    - name: r_shoulder
      struct:
      - name: x
        dtype: float32
      - name: y
        dtype: float32
    - name: l_shoulder
      struct:
      - name: x
        dtype: float32
      - name: y
        dtype: float32
  - name: bbox
    struct:
    - name: x_min
      dtype: uint16
    - name: y_min
      dtype: uint16
    - name: x_max
      dtype: uint16
    - name: y_max
      dtype: uint16
  - name: metadata
    struct:
    - name: license
      dtype: string
    - name: photo_url
      dtype: string
  splits:
  - name: train
    num_bytes: 24947710877.238
    num_examples: 241982
  - name: validation
    num_bytes: 675281167.698
    num_examples: 6531
  download_size: 25546589444
  dataset_size: 25622992044.935997
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
---