Developer Trains 125M-Parameter Transformer for On-Device Piano Autocomplete
A developer has trained a 125M-parameter transformer model to autocomplete piano performances in real time, processing about 108 notes per second on an iPhone 15. The model runs entirely on-device, allowing users to prompt it by playing a few notes on a MIDI piano, similar to code autocomplete tools like GitHub Copilot. The accompanying app is free to try.
Coverage timeline
Hacker Newssimedw
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.