Introduction
A bilingual Japanese-English ASR model that utilizes ReazonSpeech, developed by the developers of ReazonSpeech.
ReazonSpeech is an open-source dataset that contains a diverse set of natural Japanese speech, collected from terrestrial television streams. It contains more than 35,000 hours of audio.
Included Training Sets
- LibriSpeech (English)
- ReazonSpeech (Japanese)
| Datset | Number of hours | URL |
|---|---|---|
| TOTAL | 35,960 | --- |
| LibriSpeech | 960 | https://www.openslr.org/12/ |
| ReazonSpeech (all) | 35,000 | https://huggingface.co/datasets/reazon-research/reazonspeech |