README.md

metadata

language:
  - en
  - ms

Replicating Emilia on Malaysian and Singaporean Speech

Emilia, https://github.com/open-mmlab/Amphion/blob/main/preprocessors/Emilia on,

1. https://huggingface.co/datasets/malaysia-ai/crawl-youtube-malaysian-cartoons

• Originally 20.8k hours.
• Still on processing.

2. https://huggingface.co/datasets/malaysia-ai/crawl-youtube

• Originally 18.7k hours.
• 3168.8 hours after processed, 1014187 audio files.
• post cleaned to 24k and 44k samples rate at https://huggingface.co/datasets/mesolitica/Malaysian-Emilia-annotated

3. https://huggingface.co/datasets/malaysia-ai/crawl-youtube-malaysian-podcast

• Originally 2.2k hours.
• 622.8 hours after processed, 213164 audio files.
• post cleaned to 24k and 44k sample rates at https://huggingface.co/datasets/mesolitica/Malaysian-Emilia-annotated

4. https://huggingface.co/datasets/malaysia-ai/crawl-youtube-singaporean-podcast

• Originally 1.2k hours.
• 175.9 hours after processed, 61180 audio files.
• post cleaning to 24k and 44k sample rates at https://huggingface.co/datasets/mesolitica/Malaysian-Emilia-annotated

5. https://huggingface.co/datasets/malaysia-ai/crawl-youtube-malaysia-parliament

• Originally 3.9k hours.
• Still on processing.

All source code at https://github.com/mesolitica/malaysian-dataset/tree/master/text-to-speech/emilia