Visualization upgrade + enable warnings

Changes

• Introduces different plotting.
• Possibility to enable warnings via enable_warnings().

Better naming and removing of unnecessary class attributes + suppress warnings

Applied changes:

• Updates the naming of some variables, and removed other unused/unnecessary variables.

• Introduces suppress_warnings() to disable warnings globally.

Patches CNN generator and updates README

The CNN generator had hard coded input dimensionality based on the example. This is patched by adding function arguments image_height and image_width.

User improvement and patch classification problem

UX: Improving Warning Messages for Improper Use

I have focused on enhancing the user experience by providing more helpful warning messages when the tool is used improperly.

Classification Boolean Variable to Correctly Assess Neural Network Designs

I introduced the classification variable (defaulted to True) in the constructor of the PytorchWrapper class. Calculating the overall accuracy of a model only makes sense in classification problems and not in regression analysis. This update ensures a more accurate assessment of the type of neural network being designed.

Additional Changes

I have cleaned up unnecessary comments and improved the descriptions of functions and classses for better clarity and maintainability.

Improvement of example in README

This release adds the necessary library imports to the example given in the README to facilitate use.

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.