VITS_Japanese
1.0.0
(Sugestão) python == 3.7
git clone https://github.com/CjangCjengh/vits.gitpip install -r requirements.txt"n_speakers" deve ser 0 em config.json
path/to/XXX.wav|transcript
dataset/001.wav|こんにちは。
O ID do alto -falante deve começar de 0
path/to/XXX.wav|speaker id|transcript
dataset/001.wav|0|こんにちは。
Se você fez isso, defina "LIMPED_TEXT" como true em config.json
# Single speaker
python preprocess.py --text_index 1 --filelists path/to/filelist_train.txt path/to/filelist_val.txt
# Mutiple speakers
python preprocess.py --text_index 2 --filelists path/to/filelist_train.txt path/to/filelist_val.txt cd monotonic_align
python setup.py build_ext --inplace
cd .. # Single speaker
python train.py -c < config > -m < folder >
# Mutiple speakers
python train_ms.py -c < config > -m < folder > Veja inference.ipynb
Veja Moegoe
docker run -itd --gpus all --name " Container name " -e NVIDIA_DRIVER_CAPABILITIES=compute,utility -e NVIDIA_VISIBLE_DEVICES=all " Image name "