Creating an AI Musician with JavaScript
summary
Generates drum patterns in-browser using Magenta and TensorFlow.js plus Tone.js and a pretrained MusicRNN checkpoint, converts note sequences to tensors, and calls continueSequence to synthesize novel beats. A live demo maps microphone applause amplitude to the model temperature (0–2) so louder claps produce higher-entropy, less predictable riffs. Outputs are exported to GarageBand or layered into a full song; demos reliably produce usable drumbeats though the neural step feels like a controllable black box.
AI-written summary · about
chapters (5)
related talks
28mMusic and AI in the Browser with TensorFlow js and Magenta js
21mMusical Deep Neural Networks in the Browser
31mBeats in the Browser - Coding Music with JavaScript
47mUnleashing your musical creativity with Magenta.js
36mMusic and Art with ML
03mPowered by TensorFlow: Creating a custom, machine learning-powered drumming arm
01h 04mKeynote 3. Interactive generative models for music
46mMusic Generation with Magenta: Using Machine Learning in