torchtraining.functional.loss module¶
-
torchtraining.functional.loss.
binary_focal_loss
(outputs: torch.Tensor, targets: torch.Tensor, gamma: float, weight=None, pos_weight=None, reduction: Callable[torch.Tensor, torch.Tensor] = None) → torch.Tensor[source]¶ See
torchtrainingnal.loss.BinaryFocalLoss
for details.
-
torchtraining.functional.loss.
multiclass_focal_loss
(outputs: torch.Tensor, targets: torch.Tensor, gamma: float, weight=None, ignore_index=- 100, reduction: Callable[torch.Tensor, torch.Tensor] = None) → torch.Tensor[source]¶ See
torchtrainingnal.loss.MulticlassFocalLoss
for details.
-
torchtraining.functional.loss.
quadruplet
(anchor: torch.Tensor, positive: torch.Tensor, negative: torch.Tensor, negative2: torch.Tensor, alpha1: float = 1.0, alpha2: float = 0.5, metric: Callable[[torch.Tensor, torch.Tensor], torch.Tensor] = <function pairwise_distance>, weight=None, reduction: Callable[torch.Tensor, torch.Tensor] = None) → torch.Tensor[source]¶ See
torchtrainingnal.loss.Quadruplet
for details.
-
torchtraining.functional.loss.
smooth_binary_cross_entropy
(outputs: torch.Tensor, targets: torch.Tensor, alpha: float, weight=None, pos_weight=None, reduction: Callable[torch.Tensor, torch.Tensor] = None) → torch.Tensor[source]¶ See
torchtrainingnal.loss.SmoothBinaryCrossEntropy
for details.
-
torchtraining.functional.loss.
smooth_cross_entropy
(outputs: torch.Tensor, targets: torch.Tensor, alpha: float, weight=None, ignore_index: int = - 100, reduction: Callable[torch.Tensor, torch.Tensor] = None) → torch.Tensor[source]¶ See
torchtrainingnal.loss.SmoothCrossEntropy
for details.