Coqui_Xtts_Streamlit_Demo

This is a simple tutorial for Coqui XTTS using Streamlit.

# miniconda run
conda create -n coqui python=3.10
conda activate coqui

# according to requirements.txt
pip install TTS
conda install ffmpeg==6.1.1
pip install streamlit==1.38.0 pydub==0.25.1 ipykernel==6.29.5

# download xtts_v2.0.2 (not v2.0.3 if you mainly utilize the model for non-English speech)
from TTS.api import TTS
TTS("xtts_v2.0.2", gpu=False)

# move it from the default download folder to your working directory
!mv [DEFAULT_DIR]/tts_models--multilingual--multi-dataset--xtts_v2.0.2/* [YOUR_WORKING_DIR]/tts_models--multilingual--multi-dataset--xtts_v2.0.2/

Run streamlit run demo.py

Reference

• https://github.com/coqui-ai/TTS
• https://huggingface.co/spaces/coqui/xtts/blob/main/app.py

Sample Screenshot