Example application using ONNX Runtime Web & Progressive Web Application (PWA) principles to deploy an OCTA image grade classifier CNN in-browser, on client devices, with offline functionality.
- Clone this repo, and on the command line, change directory to
onnx-pwa-demo/ - Install Node Version Manager from this link
- Install node package manager:
nvm install --lts - Run the setup script:
npm run setup - Run the local dev server:
netlify dev - (Hint) if you want to use webgpu, ensure that you are running a supported browser
- Please follow this demo
- Host you model on huggingface.
- Go the
public/models.jsand add a new model. Use the existing model as a template. You will need to write your own pre/post processing functions - Write a new model test file based on for your new model
public/super_resolution-test.spec.jsto ensure your model is working correctly - (Hint) if your model doesn't work, it's probably the pre and post processing functions double check those.