metadata

language:
  - ps
license: mit
pretty_name: Pashto Synthetic Speech Dataset (2k)
size_categories:
  - 1K<n<10K
tags:
  - audio
  - speech
  - pashto
  - tts
  - synthetic
task_categories:
  - automatic-speech-recognition
metrics:
  - wer
paperswithcode:
  eval_results: []
  preferred_citation: |-
    @dataset{pashto_synthetic_speech_2k,
      title = {Pashto Synthetic Speech Dataset (2k)},
      year = {2023}
    }

Pashto Synthetic Speech Dataset (2k)

This dataset contains 4000 synthetic speech recordings in the Pashto language, with 2000 male voice recordings and 2000 female voice recordings.

Dataset Information

Dataset Size: 2000 sentences
Total Recordings: 4000 audio files (2000 male + 2000 female)
Audio Format: WAV, 44.1kHz, 16-bit PCM
Sampling Rate: 44.1kHz (44100 Hz)
Voices:
- Male: ps-AF-GulNawazNeural
- Female: ps-AF-LatifaNeural
Total Audio Duration:
- Male: 0.00 seconds
- Female: 0.00 seconds
Average Clip Duration:
- Male: 0.00 seconds
- Female: 0.00 seconds

Dataset Structure

pashto_speech_2k/
  ├── audio/
  │   ├── male/        # Male voice recordings
  │   └── female/      # Female voice recordings
  ├── manifest.json    # Combined dataset manifest
  ├── male_manifest.json     # Male voice manifest
  ├── female_manifest.json   # Female voice manifest
  ├── dataset_info.json      # Dataset metadata
  └── README.md        # This file

Usage

This dataset can be used for:

Training ASR (Automatic Speech Recognition) systems for Pashto
Text To Speech (TTS) systesm for Pashto
Voice activity detection
Speaker verification
Speech synthesis evaluation

License

This dataset is licensed under the MIT License.

Citation

If you use this dataset in your research, please cite:

@dataset{pashto_synthetic_speech_2k,
  title = {Pashto Synthetic Speech Dataset (2k)},
  author = {Hanif Rahman},
  year = {2025},
  publisher = {HuggingFace},
  howpublished = {\url{https://huggingface.co/datasets/ihanif/pashto_speech_2k}}
}

Acknowledgments

This dataset was created using Microsoft Edge TTS voices.