ML/AI models in JavaScript environment
summary
Live demo of running ONNX models in the browser and in Electron using onnxruntime-web, with source code on GitHub and slides linked. Shows an image classifier and a small recommendation-style model performing end-to-end inference plus postprocessing, using WASM and WebGL backends so computation runs client-side. Compares runtime behavior across environments, points out integration challenges (model loading, memory, threading) and proves the demo works interactively with usable latency and real output.
AI-written summary · about
chapters (5)
- 0:00Introduction to ML in JS
- 5:49ML models in production
- 13:10Electron apps with ML
- 20:37Inference runtime challenges
- 26:23Performance optimization
related talks
33mMachine Learning in the Browser: Fast Iteration with ONNX & WebAssembly
51mMaking neural networks run in browser with ONNX
19mML at Client Side
41mEscaping the Cloud: High-Performance AI in your Browser
33mBringing AI to the Heterogeneous Edge with WebAssembly & ONNX
15mThe future of machine learning using JavaScript
05mONNX Runtime
23mNode.js = Machine Learning