---
dataset_info:
  features:
  - name: speaker_id
    dtype: string
  - name: file_id
    dtype: string
  - name: audio
    dtype: audio
  - name: label
    dtype: string
  splits:
  - name: train
    num_bytes: 752441988.29
    num_examples: 6261
  download_size: 718959653
  dataset_size: 752441988.29
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
pretty_name: NepDS, ILPRL, Kathmandu University
---


# Nep_DS Dataset

The **Nep_DS dataset** is a collection of Nepali speech data, including **6,262** audio recording files in WAV format downsampled to **16KHz** and a corresponding TSV file, **"Txt_Ref.tsv,"** containing audio file names (i.e., ID) and their transcriptions (text).

## Contents

- **wavs/**: Folder with downsampled audio files in WAV format (16KHz).  
- **Txt_Ref.tsv**: TSV file mapping audio file names to their transcriptions.  

## Usage

The dataset is suitable for various speech-related tasks, such as:  

- Speech recognition  
- Speaker identification  
- Language modeling  
- Speech synthesis  
- Acoustic analysis  

## Citation

If you use the **Nep_DS dataset**, please cite it accordingly:  

```bibtex
@dataset{Nep_DS,
  author = {Shishir Paudel and Bal Krishna Bal},
  title = {Nep_DS Dataset},
  year = {2022},
  url = {https://ilprl.ku.edu.np/}
}