Show HN: Deploy Pytorch to Core ML tutorial: live and trainable spectrograms (ml-illustrated.com)

Part two of my pandemic project for deploying Pytorch to iOS via Core ML, this tutorial consisting of a model with trainable spectrogram layers for analyzing live audio captured via microphones. The trickiest part is the Core ML conversion step, which is validated via unit tests as a starting point for more complex ML-enabled iOS apps.

Any feedback on improving the tutorial and the app would be appreciated!