README.md

---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: metadata
    struct:
    - name: frame
      dtype: string
    - name: track_ids
      sequence: int32
    - name: class_ids
      sequence: int32
    - name: class_names
      sequence: string
    - name: bboxes
      dtype:
        array2_d:
          shape:
          - null
          - 4
          dtype: float32
  splits:
  - name: train
    num_bytes: 27332393.0
    num_examples: 726
  download_size: 27177436
  dataset_size: 27332393.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---