Sample rate: 24kHz
Steps: 330,000
Pretrained model: None
3 hours and 10 minutes of audio data
⋅⋅* Combination of english, tagalog and japanese singing audio data
⋅⋅* Speech audio data of english and tagalog
⋅⋅* UTAU audio samples of C3 and C4
I do not use this as a pretrain model because it's often unstable
**I do not suggest to use this as a pretrain model because it's often unstable.
But if you are going to use this for pretraining, you need to take note of these:
- Properly credit this model and the creator of the model because the audio date came from my own voice.
- Do not** release** a model that was pretrained using this model with sombody's voice (celebrities, politicians, strangers, friends, etc) without the voice providers' consent/permission.
Credit me if you use him because the audio data came from my own voice