Releases: sovaai/sova-dataset
Releases · sovaai/sova-dataset
Release v0.4.0
Added Chinese datasets (3 475,1 hours, 321 Gb):
ZhYoutube contains folders:
- zh-CN – China
- zh-SG – Singapore
- zh-TW – Taiwan
- zh-Hans – Simplified hieroglyphs
- zh-Hant – Traditional hieroglyphs
- zh – Without meta information
Release v0.3.0
- Added new datasets (17 451,06 hours, 1,83 TB):
- RuYoutube (RU)
- New cloud hosting for datasets. Legacy links are no longer supported.
Release v0.2.0
- Added new datasets (~11 402 hours, ~1,1 TB):
- EngAudiobooksOriginal (EN)
- EngAudiobooksNoisy (EN)
- RuAudiobooksDevices (RU)
- RuDevices (RU)
- New cloud hosting for datasets. Legacy links are no longer supported.