EMCqMRI.core.engine.train_model module

class EMCqMRI.core.engine.train_model.Trainer(device=[], max_epochs=[], train_data_loader=[], network=[], optimizer=[], loss_function=[], prepare_batch=[], log_training_fun=[], config_object=[])

Bases: object

Performs a forward pass through a given inference model.

Depending on the options set, it might save intermediate results and training checkpoints and/or display the intermediate results.

Parameters

• ([type (config_object) – Configuration]): [Object containing all backend configuration settings]

• config_object.args (Required) –

  • epochs

  • inference_model

  • dataloader

  • signal_model (if config_object.args.inference_model.__require_initial_guess__ == True)

  • numberOfPatches

  • objective_fun

  • optimizer

  • saveResults

  • saveResultsPath

  • saveCheckpoint

  • saveCheckpointPath

  • usePatchesAsBatches

run()