LEARN

Tutorials

Tutorials

A hands-on series of Jupyter notebooks that builds from first principles to production patterns. Each notebook is self-contained and runnable; work through them in order for a guided introduction, or jump to a topic you need.

Prerequisites: Python 3.12+. A few later notebooks import torch.nn, so install PyTorch to run those cells, or just read along.

#TutorialWhat you'll learn
1Why LacoThe problem Laco solves and the lie-typing idea
2First Stepscall, partial, instantiate: the core loop
3Lazy Call & PartialDeferred construction and partial application
4Hyperparameters & InterpolationParameters, references, and ${...} interpolation
5Loading, Saving & the CLIRound-tripping configs and overrides from the command line
6Nested Configs & ContainersComposing configs and working with containers
7Typed Groups & SchemasGroup placeholders and schema-checked composition
8Pipeline ConfigsMulti-stage pipelines as configuration
9Tasks & the App LoopWiring configs into a runnable application
10TracingInspecting how a config resolves
11Production PatternsReproducibility and scaling to real experiments

New here? Start with Why Laco, then First Steps.