ESPNet TTS with Streamlit GUI
MIT License
A Streamlit GUI interface for ESPnet2 + TTS(Text-to-speech).
https://user-images.githubusercontent.com/18676/230901038-f9351513-1d68-48ca-a6a0-e1a367e5a0eb.mp4
Code is based on ESPnet2 TTS Colab demo
https://espnet.github.io/espnet/notebook/espnet2_tts_realtime_demo.html
Currently only Japanese TTS is supported.
typeguard
version used in ESPNet2))$ python -m pip install -r requirements.txt
$ streamlit run espnet_tts_app_streamlit.py
Install pyinstaller,
$ python -m pip install pyinstaller
Then
$ pyinstaller espnet-tts-app.spec
There is no easy way to quit/terminate Streamlit app... (Especially on Windows)
Please first do ctrl-c
in Terminal window to terminate streamlit process, then close the browser window.
espnet-tts-streamlit
source code is licensed under MIT license.
pretrained model and audio datasets have their own licenses and terms of use.
kan-bayashi/tsukuyomi_full_band_vits_prosody
tsukuyomi_full_band_vits_prosody
model was trained using Tsukuyomi-chan copus and JUST copus(CC-BY-SA 4.0), thus you cannot use this pretrained model commercially. Use it for research purpose and personal use. DYOR.EoL.