voicemd

Voice classification

Setup

Clone the repo and install the project dependencies. It is recommended to use a virtual environment (e.g. conda)

git clone https://github.com/jerpint/voicemd
cd voicemd
pip install -e .

Make a prediction

From script

First, download the weights locally:

cd ~/voicemd/voicemd
curl -L -o model.pt "https://www.dropbox.com/s/r3ydpfxr3h4vlzg/best_model_commonvoice?dl=1"

Then run as a script:

from voicemd.predict import make_a_prediction

make_a_prediction(
    sound_filepath='/path/to/your/sound.wav',
    config_filepath='config.yaml',
    best_model_path='model.pt'
)

Online Demo (Colab)

• Small dataset: Use the demo colab here
• Commonvoice dataset: Use the demo colab here