A simple package to quickly asses the quality of a neural network design

The 'simple-pytorch-wrapper' package is a lightweight PyTorch wrapper that can be used to fasten process of training and setting up arbitrary Neural Network to quickly test an idea/setup. The wrapper provides an interface for both standard neural networks and CNNs, but can be extended to any architecture, with built-in visualization and performance tracking capabilities. The wrapper is customizable and aims to be used on any dataset.