LongTimeScope / README.md
ruili0's picture
Initial dataset upload with QA and OCR types
fee8075 verified
|
raw
history blame
674 Bytes
metadata
dataset_info:
  features:
    - name: index
      dtype: int64
    - name: video
      dtype: string
    - name: input
      dtype: string
    - name: length
      dtype: int64
    - name: replace_time
      dtype: float64
    - name: type
      dtype: string
    - name: answer
      dtype: string
  splits:
    - name: OCR
      num_bytes: 88981
      num_examples: 150
    - name: QA
      num_bytes: 134773
      num_examples: 300
    - name: temporal
      num_bytes: 189148
      num_examples: 450
  download_size: 68730
  dataset_size: 412902
configs:
  - config_name: default
    data_files:
      - split: OCR
        path: data/OCR-*
      - split: QA
        path: data/QA-*
      - split: temporal
        path: data/temporal-*