Tutorials & Examples

Step-by-step guides covering common ShiftKit workflows.

---

#	Tutorial	What you'll learn
1	MNIST → Noisy MNIST: Source-Only vs MMD vs DANN	End-to-end DA experiment comparing three methods on the built-in benchmark
2	MMD vs LMMD comparison	How class-conditional alignment improves over global MMD
3	Swapping the model	Using MLP instead of CNN; bring-your-own architecture
4	Registering a custom dataset	Plug in any PyTorch dataset pair without modifying the library
5	Tuning the DA weight λ	Systematic λ sweep and how to read the results
6	Regression with Domain Adaptation	Sine wave phase shift and California Housing geographic split using MMDRegressionTrainer
7	Node-level DA with PyTorch Geometric	Graph neural networks, node-level domain adaptation, and PyTorch Geometric integration
8	SIDDA Regression & Latent-Space Reweighting	SIDDARegressionTrainer on 2D blobs; comparing plain SIDDA, CE-reweight, OT-reweight, and both