mixnet.utils.DataLoader
View source on GitHubTable of contents
DataLoader class
Loading the preprocessed data of subject-dependent and subject-independent setting
mixnet.utils.DataLoader()
Arguments:
| Arguments | Description | Default |
|---|---|---|
| dataset | str dataset name (‘BCIC2a’, ‘BCIC2b’, ‘BNCI2015_001’, ‘SMR_BCI’, ‘HighGamma’, ‘OpenBMI’). | |
| train_type | str traing type (‘subject_dependent’, ‘subject_independent’). | None |
| data_type | str traing type (‘fbcsp’, ‘spectral_spatial’, ‘time_domain’, ‘spectral_spatial_signals’). | None |
| num_class | int number of classes | 2 |
| subject | int subject ID. (start from 1) | None |
| data_format | str data format None = not change,‘NCTD’=(#n_trial, #channels, #time, #depth), ‘NTCD’=(#n_trial, #time, #channels, #depth), ‘NSHWD’=(#n_trial, #n_subbands, #height, #width, #depth) | None |
| dataset_path | str path of the processed data | ‘/datasets’ |
| **kwargs |
load_train_set method
Loading the training set
DataLoader.load_train_set(fold, **kwargs)
Arguments:
| Arguments | Description | Default |
|---|---|---|
| fold | int fold number. (start from 1) | |
| **kwargs |
Returns: X, y
load_val_set method
Loading the validation set
DataLoader.load_val_set(fold, **kwargs)
Arguments:
| Arguments | Description | Default |
|---|---|---|
| fold | int fold number. (start from 1) | |
| **kwargs |
Returns: X, y
load_test_set method
Loading the test set
DataLoader.load_test_set(fold, **kwargs)
Arguments:
| Arguments | Description | Default |
|---|---|---|
| fold | int fold number. (start from 1) | |
| **kwargs |
Returns: X, y
Example 1: DataLoader for MIN2Net
from mixnet.utils import DataLoader
loader = DataLoader(dataset='OpenBMI',
train_type='subject_independent',
subject=1,
data_format='NTCD',
data_type='time_domain',
dataset_path='/datasets')
X_train, y_train = loader.load_train_set(fold=1)
X_val, y_val = loader.load_val_set(fold=1)
X_test, y_test = loader.load_test_set(fold=1)
Example 2: DataLoader for MixNet
from mixnet.utils import DataLoader
loader = DataLoader(dataset='OpenBMI',
train_type='subject_independent',
subject=1,
data_format='NTCD',
data_type='spectral_spatial_signals',
dataset_path='/datasets')
X_train, y_train = loader.load_train_set(fold=1)
X_val, y_val = loader.load_val_set(fold=1)
X_test, y_test = loader.load_test_set(fold=1)
Example 3: DataLoader for EEGNet and DeepConvNet
from mixnet.utils import DataLoader
loader = DataLoader(dataset='OpenBMI',
train_type='subject_independent',
subject=1,
data_format='NDCT',
data_type='time_domain',
dataset_path='/datasets')
X_train, y_train = loader.load_train_set(fold=1)
X_val, y_val = loader.load_val_set(fold=1)
X_test, y_test = loader.load_test_set(fold=1)