A simple neural network library written in V
This is a simple personal project, helping me understand how neural networks work by building one.
THIS LIBRARY IS NOT INTENDED FOR PRODUCTION USE
- Simple neural networks taking in 1D f64 tensors as inputs, with as many layers as you want
- sigmoid, relu, and tanh ativation functions
- square-mean and logloss loss functions
- simple gradient descent backpropagation algorithm
- the ability to save a network as a zip file to train it, and later reuse it