Datasets:

alavanaico
/

Alavan-TTS-dataset-Setare

pretty_name: Alavan TTS — Setare
tags:
  - tts
  - text-to-speech
  - persian
task_categories:
  - text-to-speech
language:
  - fa
license: cc-by-nc-4.0
dataset_info:
  features:
    - name: audio
      dtype:
        audio:
          sampling_rate: 22050
    - name: text
      dtype: string
  splits:
    - name: train
      num_bytes: 785898839
      num_examples: 991
  download_size: 749253823
  dataset_size: 785898839
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

Alavan TTS — Setare

Persian text-to-speech dataset with paired audio and text.

Dataset Summary

This dataset contains Persian speech recordings with aligned transcriptions. It is designed for training and evaluating TTS (Text-to-Speech) systems.

Dataset Structure

data/: folder containing the .wav audio files
MetaData.csv: CSV file with two columns:
- path: relative path to each audio file (e.g., data/0001.wav)
- text: Persian transcription of the corresponding audio

Example

Text: استفاده از موقعیت‌ها و به موقع گل زدن خیلی در این بازی مهم است.

Website

https://alavan.co.ir

Citation

@dataset{alavan_setare_tts,
  title = {Alavan TTS — Setare},
  author = {AlavanAi},
  year = {2025},
  url = {https://huggingface.co/datasets/alavanaico/Alavan-TTS-dataset-Setare}
}