mypt is short for My Pytorch. This is a simple library implementing routines tasks and subroutines often used in Deep Learning tasks.
The library is under construction. As of the moment of writing this readme file, the following sections are implemented:
- backbones: code to faciliate working and manipulating famous Feature Extractors such as Resnet and Alexnet
- code_utilities: utility code such as file manipulation for visual tasks, ensuring reproducibility...
- dimension_analysis: code to determine the shape of the output when passed through a given torch.nn.Module without specifically calling the module.
- data: dataset classes for specific task and automating data-related code (e.g initializing a reproducible dataloader )
- subroutines: K-nearest neighbors with Pytorch