The demo was initially YOLOv3 behind a Flask server on GPU, and the Vue frontend hitting the backend for predictions. I wanted to try everything in the browser, so using CocoSSD instead and did a lot of refactoring to actually isolate logic in multiple components, but the inference is really slow on my computer...so right now I'm putting this project on pause.
Detect a maximum of objects the model can detect, in a minimum of time.
npm install
npm run serve
npm run build
npm run lint