blues-lab/smart-speaker-implementation-demos

Blues Smart Speakers Demo

Running the Notebooks

The smark speaker demo shows Noise Reduction, Audio Fingerprinting, Wake word detection, Vad detection, and Text Classification

To run the fingerprint part of the demo, run the docker-compose up within the dejavu_repo. Then add some mp3 files within the dejavu_repo/mp3. Then use the docker_id with it. The one shown during the demo was "Field of Hopes and Dreams Remix". A good way to test quickly is via youtube-dl

To run the general classifier, check out the README.md within general_classifier and download the required model files

There are 2 notebooks that you can check out. the smart_speaker demo and the classifier notebook.

For link to the slides, https://docs.google.com/presentation/d/1X3Brj00kGBwNq0OSHT47DVG3nRY0FQzISy2TIan0GrI/edit?usp=sharing