A simple JavaScript neural network implementation with built-in backpropagation using an automatic differentiation engine similar to torch.autograd.
The goal of the project is to understand how the backpropagation mechanism works in automatic mode for a neural network of any complexity.
Start the local web server from the command line from the project directory:
python3 -m http.server 8000
and then open the page http://localhost:8000/test.html in your browser.