Greedy training =============== **Method.** Greedy ROM training adds full-order samples where the current ROM has the largest estimated error. The model is rebuilt until a stopping criterion is reached. **Example.** The notebook demonstrates iterative sample selection, ROM rebuilding, and error monitoring. Notebook: :doc:`notebooks/greedy_training`. For algorithm details, settings, and full source listings, see :doc:`greedy_training_details`. .. toctree:: :hidden: greedy_training_details