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 |