This repository provides a best-practice template for organizing deep learning projects. It wraps common deep learning tasks such as model building, training loops, and configuration management to reduce boilerplate code. The template leverages PyTorch along with high-level libraries like ignite for compact training loops and yacs for configuration handling.