v0.0.1: Pre-trained model

This model contains only few layers. It was trained that using:

• CSTR VCTK Corpus (Reduced to about 8 files per speaker)
• Common Voice (Japanese) (Balanced with gender tag before preprocess, about 600 speakers)

v0.0: Pretrained model

v0.0.0

add .gitkeep for keep data structures