trinhtuanvubk/bark-voice-cloning

Personal customization of some bark-voice-cloning implementations

Bark-Voice-Cloning

Repository Description

This repository is a derivative of some other Bark-Cloning implementations. It builds upon their work and incorporates additional features and modifications specific to this project.

Installing

• Conda:

conda create -n bark-env python=3.10.10
conda activate bark-env
pip install -r requirements.txt

• Docker:

docker build -t bark-voice-clone .
docker run -t --gpus all --name bark-env --restart always bark-voice-clone:latest
docker exec -itd bark-env bash
pip install -r requirements.txt 

Pineline

• To generate npz:

python3 main.py --scenario generate_npz

• To clone voice:

python3 main.py --scenario clone_voice

Todo

• Prepare Vietnamese data
• Traning Hubert with Vietnamese data
• Modify params for beter results