torch_base

Why we included this project

Most PyTorch users have a personal stash of boilerplate they copy between projects. TorchBase packages that common scaffolding into a small skeleton, with a directory layout that separates model code and data, and gives training scripts their own place. It also includes small utilities for logging and checkpointing, plus a simple way to visualize results. This is not a framework you learn; it is a template you read and copy, then trim to fit. The code is short enough to follow in an afternoon, so you can see exactly what each piece does and delete what you do not need. Treat it as a starting point for your own workflow rather than a dependency you rely on.