Trained models based on https://github.com/egorsmkv/flowtron
git clone https://github.com/egorsmkv/flowtron
cd flowtron
# Next, read NOTES.txt
Note:
- https://github.com/egorsmkv/flowtron is a version with phonemes
- https://github.com/egorsmkv/flowtron-uk-graphemes is a version with graphemes
Download here: https://github.com/egorsmkv/ukrainian-flowtron-tts/releases/download/v0.1-alpha/demo_kotiki.wav
- Download a checkpoint
Look here: https://github.com/egorsmkv/ukrainian-flowtron-tts/releases
- Do inference
python3 inference.py --eng 0 --sigma 0.5 -c config.json -f model_101000.pt -w waveglow_256channels_ljs_v3.pt -t "Вітаю вас, мене звати Лада" -i 0
- Ukrainian RADTTS: https://github.com/egorsmkv/ukrainian-radtts