Audio Event Detection via Deep Learning

Robert Coop, Ph.D.

This is the companion code for my talk at the Data Science ATL Conference 2018.

In order to run --

1. Clone this repository into a new virtual environment.
2. Run pip install -r requirements.txt to get the required packages
3. Download https://storage.googleapis.com/audioset/vggish_model.ckpt and https://storage.googleapis.com/audioset/vggish_pca_params.npz and place them in the dsatl_lib directory.
4. Open up Jupyter Notebook and view the code.

NOTE - you should use a virtual environment and pip install -r requirements.txt to run this code. In particular, this code was built off PyDub 0.20, which has some odd incompatibility with PyDub 0.23. There may be other similar issues. You've been warned.