/neural-network-fs

Creating a neural network from scratch without the use of libraries like TensorFlow, Keras, PyTorch, etc. Only using python math libraries.

Primary LanguagePython

A Neural Network from scratch

For experimental purposes. This is mostly just to observe the math behind a neural network and how each part comes together to allow the model to "learn."

image

A follow-up to this will be building a simple transformer from scratch, in a similar way. (But maybe with the use of a little more help, library wise.)