Datasets:
Reazon Speech v2 DENOISED Same Speaker Pair
prj-beatrice/reazon-speech-v2-denoised-titanet-embeddings のコサイン類似度の総和がなるべく大きくなるように litagin/reazon-speech-v2-denoised の音声をペアリングしたデータセットです。
事前に音声データの簡易的な重複除去を行ったものの、除去しきれていない重複データ同士がペアリングされてしまったケースが一定見られるため、ご注意ください。
Data Source
Generated from litagin/reazon-speech-v2-denoised, which is based on reazon-research/reazonspeech.
License
CDLA-Sharing-1.0 with Japanese Copyright Act Article 30-4 Restriction
- Downloads last month
- 3,383