Sound Event Detection with YAMNet+tw on Raspberry Pi

Installation

1. Raspberry Pi 3 Model B+ and Voice HAT microphone are good enough
2. Headless boot up and SSH in
3. Clone this repo
   git clone https://github.com/x1001000/sed-yamnet-raspberrypi
4. Install dependencies, drivers, and configs
   bash sed-yamnet-raspberrypi/install-deps.sh
5. Reboot to take effect
   sudo reboot
6. Run SED
   cd && cd sed-yamnet-raspberrypi/yamnet && python3 sed.py
7. There you go!

References

1. https://www.tensorflow.org/guide/keras/transfer_learning
2. https://www.tensorflow.org/tutorials/images/transfer_learning
3. https://www.tensorflow.org/hub/tutorials/yamnet
4. https://github.com/tensorflow/models/tree/master/research/audioset
5. https://github.com/Qengineering/Tensorflow-Raspberry-Pi/
6. tensorflow/tensorflow#44467
7. https://github.com/hellmanj/AIY-voice-kit-python/blob/master/HACKING.md
8. https://people.csail.mit.edu/hubert/pyaudio/