README.md

metadata

dataset_info:
  features:
    - name: audio
      dtype:
        audio:
          sampling_rate: 16000
    - name: sentence
      dtype: string
  splits:
    - name: train
      num_bytes: 14520520179.425
      num_examples: 76441
  download_size: 13849169266
  dataset_size: 14520520179.425
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: mit
task_categories:
  - automatic-speech-recognition
  - translation
language:
  - ja
tags:
  - japanese
  - cleaned
  - shuffeled
  - flattened
pretty_name: 3 japanese languge datasets cleaned and ready for whisper training
size_categories:
  - 10K<n<100K