unorthodox-oddball / text-generation-webui-edge-tts

A very simple implementation of edge_tts w/ RVC for oobabooga text-generation-webui.

text-generation-webui-edge-tts

A simple extension for the text-generation-webui by oobabooga that uses edge_tts for audio output. It also supports post-processing using RVC.

How to install

Assuming you already have the text-generation-webui set up (min version 1.6.1):

1. Activate the conda environment using cmd_windows.bat
2. Enter the text-generation-webui/extensions/ directory and clone this repository

cd text-generation-webui/extensions/
git clone https://github.com/Unorthodox-oddball/text-generation-webui-edge-tts.git

3. Install the requirements

pip install fairseq-0.12.3.1-cp311-cp311-win_amd64.whl
pip install -r edge_tts/requirements.txt

4. Add --extensions edge_tts to your startup script
   or
   enable it through the Session tab in the webui
5. Download the required RVC models and place them in the extensions/edge_tts/models folder

curl -L -O https://huggingface.co/lj1995/VoiceConversionWebUI/resolve/main/hubert_base.pt
curl -L -O https://huggingface.co/lj1995/VoiceConversionWebUI/resolve/main/rmvpe.pt

6. Add your .pth files to extensions/edge_tts/rvc_models

Notes

Edge TTS is a free API provided by Microsoft. An internet connection is required for the TTS to function.

RVC was inspired by rvc-tts-webui