alibi.models.pytorch.model

This module tries to provided a class wrapper to mimic the TensorFlow API of tensorflow.keras.Model. It is intended to simplify the training of a model through methods like compile, fit and evaluate which allow the user to define custom loss functions, optimizers, evaluation metrics, train a model and evaluate it. Currently it is used internally to test the functionalities for the Pytorch backend. To be discussed if the module will be exposed to the user in future versions.

Model

Inherits from: Module

Constructor

Model(self, **kwargs)

Methods

compile

compile(optimizer: torch.optim.optimizer.Optimizer, loss: Union[Callable, List[Callable]], loss_weights: Optional[List[float]] = None, metrics: Optional[List[alibi.models.pytorch.metrics.Metric]] = None)

Name	Type	Default	Description
optimizer	torch.optim.optimizer.Optimizer		Optimizer to be used.
loss	Union[Callable, List[Callable]]		Loss function to be used. Can be a list of the loss function which will be weighted and summed up to compute the total loss.
loss_weights	Optional[List[float]]	None	Weights corresponding to each loss function. Only used if the loss argument is a list.
metrics	Optional[List[alibi.models.pytorch.metrics.Metric]]	None	Metrics used to monitor the training process.

compute_loss

compute_loss(y_pred: Union[torch.Tensor, List[torch.Tensor]], y_true: Union[torch.Tensor, List[torch.Tensor]]) -> Tuple[torch.Tensor, Dict[str, float]]

Name	Type	Default	Description
y_pred	Union[torch.Tensor, List[torch.Tensor]]		Prediction labels.
y_true	Union[torch.Tensor, List[torch.Tensor]]		True labels.

Returns

• Type: Tuple[torch.Tensor, Dict[str, float]]

compute_metrics

compute_metrics(y_pred: Union[torch.Tensor, List[torch.Tensor]], y_true: Union[torch.Tensor, List[torch.Tensor]]) -> Dict[str, float]

Name	Type	Default	Description
y_pred	Union[torch.Tensor, List[torch.Tensor]]		Prediction labels.
y_true	Union[torch.Tensor, List[torch.Tensor]]		True labels.

Returns

• Type: Dict[str, float]

evaluate

evaluate(testloader: torch.utils.data.dataloader.DataLoader) -> Dict[str, float]

Name	Type	Default	Description
testloader	torch.utils.data.dataloader.DataLoader		Test dataloader.

Returns

• Type: Dict[str, float]

fit

fit(trainloader: torch.utils.data.dataloader.DataLoader, epochs: int) -> Dict[str, float]

Name	Type	Default	Description
trainloader	torch.utils.data.dataloader.DataLoader		Training data loader.
epochs	int		Number of epochs to train the model.

Returns

• Type: Dict[str, float]

load_weights

load_weights(path: str) -> None

Loads the weight of the current model.

Name	Type	Default	Description
path	str

Returns

• Type: None

save_weights

save_weights(path: str) -> None

Save the weight of the current model.

Name	Type	Default	Description
path	str

Returns

• Type: None

test_step

test_step(x: torch.Tensor, y: Union[torch.Tensor, List[torch.Tensor]])

Name	Type	Default	Description
x	torch.Tensor		Input tensor.
y	Union[torch.Tensor, List[torch.Tensor]]		Label tensor.

train_step

train_step(x: torch.Tensor, y: Union[torch.Tensor, List[torch.Tensor]]) -> Dict[str, float]

Name	Type	Default	Description
x	torch.Tensor		Input tensor.
y	Union[torch.Tensor, List[torch.Tensor]]		Label tensor.

Returns

• Type: Dict[str, float]

validate_prediction_labels

validate_prediction_labels(y_pred: Union[torch.Tensor, List[torch.Tensor]], y_true: Union[torch.Tensor, List[torch.Tensor]])

Name	Type	Default	Description
y_pred	Union[torch.Tensor, List[torch.Tensor]]		Prediction labels.
y_true	Union[torch.Tensor, List[torch.Tensor]]		True labels.