Machine Learning & JS Take your PWA to the next level
summary
Claims that you can build and train neural nets entirely in the browser using TensorFlow.js with WebGL acceleration and that a TensorFlow backend for Node.js exposes GPU/TPU on the server side. Demo shows a Tinder‑style image/gesture classifier trained from webcam images via transfer learning on MobileNet converted from Keras into a JSON model manifest, uses fit() in the browser, trains fast enough for live labeling and yields workable predictions.
AI-written summary · about
chapters (12)
- 0:00Intro
- 0:54What is Machine Learning
- 5:37Architecture
- 6:47Performance
- 7:47Demo
- 8:54Training
- 11:32Transfer Learning
- 12:14Load Model
- 13:49Collect Training Data
- 15:19Compile Model
- 16:50Fit Method
- 18:30Conclusion
related talks
32mWe Also Can Do It! Machine Learning in Javascript!
15mMachine learning for next gen web apps with TensorFlow.js
34mA Powerful Introduction to Machine Learning with Js
34mA Powerful Introduction to Machine Learning with Js
20mTensorFlow.js Bringing Machine Learning to the Web and Beyond
24mThe Future of Machine Learning & Javascript
54mDeep Learning in JS
39mMachine Learning Magic for Your JavaScript Application