Using Machine Learning Models in JavaScript Application
summary
Shows how to run and retrain models with TensorFlow.js in Node and the browser, using a tiny linear regression on simulated X/Y arrays trained for 100 epochs that predicts a five‑room house at ~3.2 (target 3.0) while loss steadily drops. Demonstrates image classification with MobileNet (cat → Egyptian/tiger/tabby probabilities; wine → 99% red wine). Notes browser backends WebAssembly/WebGL/WebGPU, Stable Diffusion dropping 30s→10s, and a Hugging Face model running faster on Node than Python, plus offline/latency/privacy tradeoffs.
AI-written summary · about
chapters (11)
- 0:00ML in JavaScript apps
- 1:04Machine learning with JS
- 6:00TensorFlow.js vs Python
- 10:35Models in the browser
- 13:45ML model types
- 15:10ML workflow
- 17:19Linear regression demo
- 22:44Image classification demo
- 26:44What's new in TensorFlow.js
- 29:15Node.js ML use cases
- 30:41JS as ML gateway
related talks
32mUsing Machine Learning Models in JavaScript Application
15mThe future of machine learning using JavaScript
21mA Journey into the World of Machine Learning with TensorFlow.js
24mTensorFlow.js + Node
41mTensorFlow.js 101: ML in the Browser and Beyond
24mThe Future of Machine Learning & Javascript
20mTensorFlow.js Bringing Machine Learning to the Web and Beyond
34mA Powerful Introduction to Machine Learning with Js