HoloNet.predicting.load_model_list

HoloNet.predicting.load_model_list(X, adj, project_name, used_target_gene_name_list=None, only_cell_type=False, model_save_folder='_tmp_save_model')

Load trained model list in model_save_folder/project_name/

Parameters

X Tensor

The feature matrix used as the input of the loading trained models.

adj Tensor

The adjacency matrix used as the input of the loading trained models.

project_name str

The name of project, such as ‘BRCA_10x_generating_all_target_gene’.

used_target_gene_name_list str | NoneOptional[str] (default: None)

If not None, can only load part of models for some given genes.

only_cell_type bool (default: False)

If true, the model only use the Feature matrix training target, serving as a baseline model.

model_save_folder str (default: '_tmp_save_model')

The father folder.

Return type

Tuple[Union[List[MGC_Model], List[List[MGC_Model]]], List[str]]

Returns

: trained_MGC_model_list_all_target :

A list of trained MGC model for generating the expression of one target gene. Or a list of the trained MGC model lists for multiple target genes, if with multiple targets.

all_target_gene_list :

A list of target gene names.