---
language:
- ms
task_categories:
- automatic-speech-recognition
- text-to-speech
---

# Pseudolabel Youtube Malay audiobooks using Whisper Large V3

Notebooks at https://github.com/mesolitica/malaysian-dataset/tree/master/speech-to-text-semisupervised/youtube-audiobook

1. Split based on 10 seconds utterances using WebRTC VAD.

## how-to

Download files,

```bash
wget https://huggingface.co/datasets/mesolitica/semisupervised-audiobook/resolve/main/bukan-kerana-aku-5secs-noisy.tar.gz
wget https://huggingface.co/datasets/mesolitica/semisupervised-audiobook/resolve/main/bukan-kerana-aku-noisy.tar.gz
wget https://huggingface.co/datasets/mesolitica/semisupervised-audiobook/resolve/main/harry-potter-5secs-noisy.tar.gz
wget https://huggingface.co/datasets/mesolitica/semisupervised-audiobook/resolve/main/harry-potter-noisy.tar.gz
wget https://huggingface.co/datasets/mesolitica/semisupervised-audiobook/resolve/main/teme-5secs-noisy.tar.gz
wget https://huggingface.co/datasets/mesolitica/semisupervised-audiobook/resolve/main/teme-noisy.tar.gz
wget https://huggingface.co/datasets/mesolitica/semisupervised-audiobook/resolve/main/semisupervised-audiobook-part1.json
wget https://huggingface.co/datasets/mesolitica/semisupervised-audiobook/resolve/main/semisupervised-audiobook-part2.json
```