Provides a new version of [Mygf-f040k] training with a mature voice Pt (beta version)

This week, we completed training and testing the pre-trained model, increasing the number of speakers from 109 to 426, integrating Chinese, Japanese, and English corpora, and using whisper data for the first time.

This model is trained on the Mygf-f040k pre-trained model, using a small portion of whispers as the dataset.

We will now provide this part of the PT model for free to share with you.

Wishing you a pleasant holiday.

Dataset: 7 min.