Mimi-Voice

Create Unmute voice embeddings

Setup

git clone git@github.com:davidbrowne17/mimi-voice.git
cd Mimi-Voice
python3.10 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

# Create the voice embeddings
python create_voice.py <input_folder> <output_folder>
# You will need access to huggingface
huggingface-cli login

Training results for the model

• MSE Loss: 0.050422
• Cosine Loss: 0.051857
• Distribution Loss: 0.001720

Full License:

You can read the full license text here:
https://creativecommons.org/licenses/by-nc/4.0

By requesting access or using this model, you agree to these terms.