---
dataset_info:
  features:
  - name: text
    dtype: string
  - name: description
    dtype: string
  - name: url
    dtype: string
  - name: short
    dtype: int64
  - name: audio_tokens
    sequence: int64
  - name: md5
    dtype: string
  - name: start_us
    sequence: float64
  - name: end_us
    sequence: float64
  - name: words
    sequence: string
  splits:
  - name: train
    num_bytes: 43164739
    num_examples: 867
  download_size: 8645119
  dataset_size: 43164739
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---

This dataset contains no audio file. The `audio_tokens` column is generated using WavTokenizer, to be used by OuteTTS 500M v0.2 model.

If you want to re-use this data to train for TTS model, download the full dataset containing WAV file from https://huggingface.co/datasets/ngxson/ThisTTS_dataset