Building Machine Learning Models on Edge with React
summary
Live browser inference using tensorflow.js, a pose‑estimation demo that overlays a skeleton on webcam feed and a sentiment classifier built with transformers.js both run entirely client‑side with no OpenAI keys. Pose demo takes about two hours to assemble and works in-browser; sentiment uses a web workers thread to offload inference to a distilled birth model so UI stays responsive. Code examples show model.load, model.detect, and simple canvas bounding boxes.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 2:37Talk overview
- 4:43JavaScript developers and ML
- 7:11TensorFlow.js ecosystem
- 10:16React ML image demos
- 12:25Building React ML classifier
- 15:36Companies using TensorFlow.js
- 17:22Other ML libraries
- 19:47OpenAI with Next.js
- 21:36Optimizing with Web Workers
- 26:52Resources and call to action
- 27:56Q&A
related talks
41mAI in Web: TensorFlow js
34mA Powerful Introduction to Machine Learning with Js
15mMachine learning for next gen web apps with TensorFlow.js
26mBuild your Machine Learning Model on Edge with React Native
19mML at Client Side
15mThe future of machine learning using JavaScript
26mMachine Learning and AI for Angular Developers
39mMachine Learning Magic for Your JavaScript Application