Рет қаралды 1,331
Almost everything in this is on the autopitch model
Model details below:
------------------------------------------------------------------------------------------------------------
Sample cleaning, labelling, preparing, etc: @PixPrucer
Training: @haru0l
Acoustic model is 77k steps @ 800+ epochs
Variance model is 77k steps @ ??? epochs (Still in training, this video uses the checkpoint at 20k steps)
This model uses PJS as a secondary speaker to enhance the lower ranges. The labeling is done by @UtaUtaUtau
No7 is attributed to No.7製作委員会, and follows a strictly no-commercial usage terms as per the database rules.
Release date TBA
------------------------------------------------------------------------------------------------------------
Song and lyrics by 岩里祐穂, 白戸佑輔
VSQx by Staircatte