PyTorch Models

PyTorch models for EEG classification.

Classes:

Name	Description
PyTorchCNN	CNN classifier implemented with PyTorch.
PyTorchLSTM	Stacked bidirectional LSTM classifier implemented with PyTorch.
PyTorchCNNLSTM	Hybrid CNN-LSTM classifier implemented with PyTorch.
PyTorchGenerator	Conditional GAN generator implemented with PyTorch.
PyTorchDiscriminator	Conditional GAN discriminator implemented with PyTorch.
PyTorchGANAugmenter	Train a PyTorch conditional GAN and append generated samples.
PyTorchTrainer	Minimal PyTorch trainer for notebook reproduction.

Functions:

Name	Description
build_classifier	Build a named PyTorch classifier.

PyTorchCNN

PyTorchCNN(config: ModelConfig | None = None)

Bases: Module

CNN classifier implemented with PyTorch.

Parameters:

Name	Type	Description	Default
config	ModelConfig | None	Optional architecture configuration.	None

PyTorchLSTM

PyTorchLSTM(config: ModelConfig | None = None)

Bases: Module

Stacked bidirectional LSTM classifier implemented with PyTorch.

Parameters:

Name	Type	Description	Default
config	ModelConfig | None	Optional architecture configuration.	None

PyTorchCNNLSTM

PyTorchCNNLSTM(config: ModelConfig | None = None)

Bases: Module

Hybrid CNN-LSTM classifier implemented with PyTorch.

Parameters:

Name	Type	Description	Default
config	ModelConfig | None	Optional architecture configuration.	None

PyTorchGenerator

PyTorchGenerator(config: ModelConfig | None = None)

Bases: Module

Conditional GAN generator implemented with PyTorch.

Parameters:

Name	Type	Description	Default
config	ModelConfig | None	Optional architecture configuration.	None

PyTorchDiscriminator

PyTorchDiscriminator(config: ModelConfig | None = None)

Bases: Module

Conditional GAN discriminator implemented with PyTorch.

Parameters:

Name	Type	Description	Default
config	ModelConfig | None	Optional architecture configuration.	None

PyTorchGANAugmenter dataclass

PyTorchGANAugmenter(
    training: TrainingConfig,
    model_config: ModelConfig | None = None,
    device: str | None = None,
)

Train a PyTorch conditional GAN and append generated samples.

The returned arrays use the same classifier layout as prepare_splits, so the augmentation can be applied before CNN, LSTM, or CNN-LSTM training.

Attributes:

Name	Type	Description
training	TrainingConfig	Training and GAN augmentation configuration.
model_config	ModelConfig | None	Optional architecture configuration.
device	str | None	Optional device string; defaults to CUDA when available, otherwise CPU.

Methods:

Name	Description
augment_training_data	Create GAN-augmented training arrays.

augment_training_data

augment_training_data(
    x_train: ndarray, y_train: ndarray
) -> GANAugmentationResult

Create GAN-augmented training arrays.

Parameters:

Name	Type	Description	Default
x_train	ndarray	Real training inputs shaped (n, time, 1, channels).	required
y_train	ndarray	One-hot real training labels.	required

Returns:

Type	Description
GANAugmentationResult	Augmented training arrays plus synthetic sample metadata.

PyTorchTrainer dataclass

PyTorchTrainer(
    training: TrainingConfig, device: str | None = None
)

Minimal PyTorch trainer for notebook reproduction.

Attributes:

Name	Type	Description
training	TrainingConfig	Training configuration controlling epochs and batch size.
device	str | None	Optional device string; defaults to CUDA when available, otherwise CPU.

Methods:

Name	Description
fit	Train a PyTorch classifier.
evaluate	Evaluate a PyTorch classifier.

fit

fit(
    model: Module,
    x_train: Any,
    y_train: Any,
    x_valid: Any | None = None,
    y_valid: Any | None = None,
) -> list[dict[str, float]]

Train a PyTorch classifier.

Parameters:

Name	Type	Description	Default
model	Module	PyTorch classifier.	required
x_train	Any	Training inputs.	required
y_train	Any	Training labels as class IDs or one-hot rows.	required
x_valid	Any | None	Reserved for API parity; currently unused.	None
y_valid	Any | None	Reserved for API parity; currently unused.	None

Returns:

Type	Description
list[dict[str, float]]	Per-epoch metrics with loss and accuracy.

evaluate

evaluate(model: Module, x: Any, y: Any) -> dict[str, float]

Evaluate a PyTorch classifier.

Parameters:

Name	Type	Description	Default
model	Module	Trained PyTorch classifier.	required
x	Any	Evaluation inputs.	required
y	Any	Evaluation labels as class IDs or one-hot rows.	required

Returns:

Type	Description
dict[str, float]	Dictionary with loss and accuracy.

build_classifier

build_classifier(
    name: str, config: ModelConfig | None = None
) -> Module

Build a named PyTorch classifier.

Parameters:

Name	Type	Description	Default
name	str	Classifier name: cnn, lstm, or cnn_lstm.	required
config	ModelConfig | None	Optional architecture configuration.	None

Returns:

Type	Description
Module	PyTorch module for the selected classifier.

Raises:

Type	Description
ValueError	If the classifier name is unknown.
ImportError	If PyTorch is not installed.