NN.Examples.DeepDives

Curated umbrella for deep-dive TorchLean examples.

These files are not beginner introductory examples. They cover cross-cutting boundaries:

• higher-order, directional, and stopped-gradient autograd transforms;
• float/runtime semantics, numerical certificates, and Arb-backed interval comparison;
• GraphSpec lowering into the training API;
• IR axis semantics and PyTorch export;
• the “one semantic universe” contract connecting execution, CROWN/IBP, and widgets;
• tensor operations, indexing, reshaping, einsum, and editor-only widget panels.

Rule of thumb for this folder: examples may contain small concrete graphs, parameters, and display data, but reusable algorithms and data structures belong in NN.* library modules. Import this umbrella when you want the whole deep-dive example surface.