Elements we need: - What is the documentation for each algorithm - How to correctly do the logging (e.g., WandB) - Good example script or template/skeleton code -- one candidate is the conditional GAN code. - How to create an evaluation script - How to handle multiple components or more complex models (e.g., VQ-GAN) -- what are best practices for this. - Hyper-parameter naming conventions. Scope: - Make a CONTRIBUTING.md file and add the above as Markdown headings with examples. - Make a `template` folder in `engiopt` source that provides a starting point.