Plug and play audio classification. It scored 96% for snoring detection as benchmark.
You can try your own audio dataset.
This repository support different types of feature extraction methods and classifiers.
Currently, k-NN and Logistic Regression are available
Currently, MLP, CNN, ResNet(Transfer Learning) are supported
Make sure to use python 3.6
pip install -r backend/requirements.txt
- Make folder called
data
with sub-folder indicates categories. - Move audio files into each sub-folders
Now, you are good to go
Simply, run python backend/src/main.py
Edit config/master_config.ini
You can select different types of feature extraction and classifier.
I recommend to set audio length to median of audio files.