Music-AI-Voices

Introduction

Music-AI-Voices is a repository of AI-generated voice models designed for creating music covers and impressions using machine learning techniques. The project encourages the legal use of sample clips and AI-generated music while discouraging piracy. The repository does not contain any datasets, only models.

Architecture

The models in this repository are built using SoftVC VITS Singing Voice Conversion (VITS SVC 4.0) and Retrieval-based Voice Conversion (RVC). These technologies enable the transformation and synthesis of singing voices through advanced machine learning algorithms.

Training

The training details for each model vary, with some models trained for thousands of epochs or steps. Each model is created by different contributors, and training methods may include using professional tools for specific vocal qualities.

Guide: Running Locally

1. Clone the Repository: Download the models from the Music-AI-Voices repository.
2. Set Up Environment: Install necessary dependencies and set up the environment for running VITS SVC or RVC models.
3. Run the Model: Follow specific instructions per model for input and execution.
4. Hardware Suggestions: For optimal performance, using cloud-based GPUs like NVIDIA A100 or similar might be beneficial for handling large model computations.

License

The repository is licensed under an "other" license, indicating specific terms that may not align with common open-source licenses. Users are advised to review the license details before using the models.