No7 (SEVEN) for DiffSinger is based on the public singing voice database sampled from voice actress Kotori Koiwai.

This model supports velocity, gender, autopitch, and also uses a finetuned vocoder for enhanced quality. (Requires the latest OpenUTAU update.)

This model has been trained together with PJS, with labeling done by UtaUtaUtau.

The terms of use of this model are identical as the TOS on the official website.

• Model range - C3-F5

• Corpus preparation, labelling, audio cleaning - PixPrucer

• Model training - haru0l