yamle.data.reconstruction module

class yamle.data.reconstruction.ECG5000ReconstructionDataModule(anomaly=False, *args, **kwargs)

Bases: ECG5000DataModule

Reconstruction data module for the ECG5000 dataset.

If anomaly is set to True, then all the anomalous cases from the training set are appended to the test set. This can be used to create an autoencoder trained only on normal cases, being unable to reconstruct the anomalous cases. Thus training an autoencoder for anomaly detection.

Parameters:

anomaly (bool) – If True, then all the anomalous cases from the training set are appended to the test set. Default: False.

outputs_dim = 2

outputs_dtype = torch.float32

inputs_dtype = torch.float32

targets_dim = 140

task = 'reconstruction'

prepare_data()

Download and prepare the data

Return type:

None

plot(tester, save_path, specific_name='')

Plots the dataset.

Return type:

None

static add_specific_args(parent_parser)

This method is used to add datamodel specific arguments to the general parser.

Return type:

ArgumentParser

available_transforms: List[str]

available_test_augmentations: List[str]

test_augmentations: List[str]