This app is a very small scaffold to get you started using Teachable Machine. It will require an internet connection to work, since it loads ml5 as well as TensorFlow from a CDN.
npm run watch
will start a webpack server at localhost:3000. From the browser you'll be able to see a simple example that recognizes "heart hands," or when you make your hands look like this
If you want to see the sound recognition or drawing recognition examples, you'll need to comment out the appropriate line in app/app.js to change which setup function is loaded.
This starter includes a script that you can use to download images off of Image Net. To download 100 random images to a directory called data/random you could run
npm run images -- -o data/randomIf you wanted instead to get 100 images with a particular wnid, then you'd execute something like this:
npm run images -- -o data/red-panda --wnid n02509815 --randomizeWhich will download 100 images of a red panda. The --randomize flag tells the downloader whether to download the first 100 images, or to scramble the list of images first. You can also use the --count flag to limit or increase the number of images downloaded. Note that many links in Image Net are broken, and you may see errors in the console while downloading.
A long time ago this was a starter on Glitch using React and Webpack. It was copied by @starakaj, and then React has been removed You can find the original at https://glitch.com/~starter-react.
This project relates to video 2 of 5 in the React Starter Kit video series.