tabensemb.model.TorchModel#

class tabensemb.model.TorchModel(*args, lightning_trainer_kwargs: Dict | None = None, **kwargs)[source]#

Bases: AbstractModel

The class for PyTorch-like models. Some abstract methods in AbstractModel are implemented.

Methods

__init__(*args, lightning_trainer_kwargs: Dict | None = None, **kwargs)[source]#

Parameters:

trainer:: A Trainer instance that contains all information and datasets and will be linked to the model base. The trainer has loaded configs and data.
program:: The name of the model base. If None, the name from _get_program_name() is used.
model_subset:: The names of models selected to be trained in the model base.
exclude_models:: The names of models that should not be trained. Only one of model_subset and exclude_models can be specified.
store_in_harddisk:: Whether to save models in the hard disk. If the global setting tabensemb.setting["low_memory"] is True, True is used.
optimizers: A dictionary of optimizer names (choose from those in torch.optim) and their hyperparameters for each model. Remember to change _initial_values() and _space() to optimize its hyperparameters.
lr_schedulers: A dictionary of lr scheduler names (choose from those in torch.optim.lr_scheduler) and their hyperparameters for each model. Remember to change _initial_values() and _space() to optimize its hyperparameters.
**kwargs:: Ignored.

`cal_feature_importance`(model_name, method[, ...])	Calculate feature importance using a specified model.
`cal_shap`(model_name[, call_general_method, ...])	Calculate SHAP values using a specified model.
`count_params`(model_name[, trainable_only])	Count the number of parameters in a `torch.nn.Module`
`get_full_name_from_required_model`(required_model)	Get the name of a required model to store or access data in `derived_tensors` passed to `AbstractNN._forward()`.
`_data_preprocess`(df, derived_data, model_name)	Perform the same preprocessing as in `_train_data_preprocess()` on a new dataset.
`_generate_dataset`(datamodule, model_name)	Generate `torch.utils.data.Dataset` for training.
`_generate_dataset_for_required_models`(df, ...)	Call `AbstractModel._data_preprocess()` to generate the dataset, output, and hidden representations for the required model
`_generate_dataset_from_tensors`(tensors, df, ...)	Perform the same preprocessing as in `_generate_dataset()` on a new dataset.
`_initial_values`(model_name)	Initial values of hyperparameters to be optimized.
`_pred_single_model`(model, X_test, verbose, ...)	Predict using the model trained in `_train_single_model()`.
`_prepare_custom_datamodule`(model_name[, ...])	Change this method if a customized preprocessing stage is needed.
`_prepare_tensors`(df, derived_data, model_name)	Transform the upcoming dataset into Tensors that has the same structures as those stored in a `tabensemb.data.datamodule.DataModule` and obtained by `tabensemb.data.datamodule.DataModule.update_dataset()`.
`_run_custom_data_module`(df, derived_data, ...)	Change this method if a customized preprocessing stage is implemented in `_prepare_custom_datamodule()`.
`_space`(model_name)	A list of `scikit-optimize` search spaces for the selected model.
`_train_data_preprocess`(model_name[, warm_start])	Processing the data from `self.trainer.datamodule` for training.
`_train_single_model`(model, model_name, ...)	`pytorch_lightning` implementation of training a pytorch model.