multiclass_metrics

MulticlassConfusionMatrix that support logit.

Classes:

Name	Description
MulticlassConfusionMatrix	MulticlassConfusionMatrix with one hot target support.
MulticlassAccuracy	MulticlassAccuracy with one hot target support.
MulticlassF1Score	MulticlassConfusionMatrix with one hot target support.

MulticlassConfusionMatrix

MulticlassConfusionMatrix with one hot target support.

Methods:

Name	Description
update	Update the metric.

Attributes:

Name	Type	Description
__module__

__module__ = 'xpdeep_utils.train.metrics_zoo.classification.confusion_matrix'

update(preds: Tensor, target: Tensor) -> None

Update the metric.

Source code in src/xpdeep/metrics/zoo/multiclass_metrics.py

def update(self, preds: Tensor, target: Tensor) -> None:
    """Update the metric."""
    target = torch.argmax(target, dim=1)
    super().update(preds, target)

MulticlassAccuracy

MulticlassAccuracy with one hot target support.

Methods:

Name	Description
update	Update the metric.

Attributes:

Name	Type	Description
__module__

__module__ = 'xpdeep_utils.train.metrics_zoo.classification.multi_class_accuracy'

update(preds: Tensor, target: Tensor) -> None

Update the metric.

Source code in src/xpdeep/metrics/zoo/multiclass_metrics.py

def update(self, preds: Tensor, target: Tensor) -> None:
    """Update the metric."""
    target = torch.argmax(target, dim=1)
    super().update(preds, target)

MulticlassF1Score

MulticlassConfusionMatrix with one hot target support.

Methods:

Name	Description
update	Update the metric.

Attributes:

Name	Type	Description
__module__

__module__ = 'xpdeep_utils.train.metrics_zoo.classification.f1_score'

update(preds: Tensor, target: Tensor) -> None

Update the metric.

Source code in src/xpdeep/metrics/zoo/multiclass_metrics.py

def update(self, preds: Tensor, target: Tensor) -> None:
    """Update the metric."""
    target = torch.argmax(target, dim=1)
    super().update(preds, target)