I use this repository to study different neural network structures and the math behind them.
This repository is organized by programming language and dataset the code is modeling.
- MNIST - handwritten digit classification
- Python/scratch_two_layer.ipynb
- a jupyter notebook that studies the math of a simple two layer neural network applied to the mnist dataset. This neural network is programmed from scratch without using tensorflow or PyTorch.
- R/scratch_conv.rmd
- an rmarkdown notebook of a simple convolutional neural network written from scratch applied to the mnist dataset.
- Python/scratch_two_layer.ipynb