Datasets:

ICML-2026
/

ToneBooks

dataset_info:
  features:
    - name: audio
      dtype: audio
    - name: text
      dtype: string
    - name: text_description
      dtype: string
    - name: voice_name
      dtype: string
  splits:
    - name: train
      num_bytes: 10469796890.924692
      num_examples: 91976
    - name: validation
      num_bytes: 551059915.075307
      num_examples: 4841
  download_size: 10547648552
  dataset_size: 11020856806
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*
license: apache-2.0
task_categories:
  - text-to-speech
  - automatic-speech-recognition
language:
  - ru
pretty_name: ToneBooks
size_categories:
  - 10K<n<100K