This is a demonstration of how to use TensorFlow.js to load a pre-trained base model, specifically MobileNet, to generate image features that can be used in transfer learning.
This web app can gather data from a user's webcam for multiple classes of data that you want to recognize. It will create and define a multi-layer perceptron that takes the image features and learns to classify new objects using them.
You can simply fork and clone this repo git clone https://github.com/ddoyediran/ml-transfer-learning.
You don't need to download any dependencies. Open the index.html in your browser with a live server e.g Live Server extension from VS code
Keep an eye on the repo, this app will be deployed soon so you can easily test on your browser without cloning