modnet.model_presets.presets module

This submodule defines some heuristic presets to be used as the first step of hyperparameter optimisation.

modnet.model_presets.presets.gen_presets(n_feat, n_samples, classification=False)

Generates sensible preset architectures and learning parameters based on number of samples and features.

Parameters:

• n_feat (int) – The number of training features available to the model.

• n_samples (int) – The number of training samples available to the model.

• classification (bool) –

Returns:

List of dictionaries to individually pass as kwargs to model.fit(...).

Return type:

List[Dict[str, Any]]