The Learning and Inference ExamplE has now arrived!
Over the past couple weeks I've been developing LIEE DIFF-SVC AI as a way to learn and teach others about DIFF-SVC in my documentation. Now, LIEE DIFF-SVC AI is available for pre-training and example purposes.
LIEE is a Diff-SVC AI model created for research purposes, voiced by a consenting voice provider (me), sourced from singing and speaking samples.
LIEE DIFF-SVC AI is now available in the Colab training notebook.
LIEE DIFF-SVC AI v1 is trained with the 24kHz vocoder.
LIEE DIFF-SVC AI v2 is trained with the 44.1kHz vocoder.
If you use them for pre-training please be sure to credit me.
You can access the voice model files and see other releases here: [ Ссылка ]
MAKE YOUR OWN DIFF-SVC VOICEBANK: [ Ссылка ]
【FOLLOW ME】
Download my UTAUs: [ Ссылка ]
Twitch: [ Ссылка ]
UTAU Twitter: [ Ссылка ]
Main Twitter: [ Ссылка ]
Soundcloud: [ Ссылка ]
![](https://i.ytimg.com/vi/R5zUgKhFn88/maxresdefault.jpg)