dacapo.experiments.tasks.losses

Submodules

• dacapo.experiments.tasks.losses.affinities_loss
• dacapo.experiments.tasks.losses.dummy_loss
• dacapo.experiments.tasks.losses.hot_distance_loss
• dacapo.experiments.tasks.losses.loss
• dacapo.experiments.tasks.losses.mse_loss

Package Contents

Classes

DummyLoss	A class representing a dummy loss function that calculates the absolute difference between each prediction and target.
MSELoss	A class used to represent the Mean Square Error Loss function (MSELoss).
Loss	Helper class that provides a standard way to create an ABC using
AffinitiesLoss	Helper class that provides a standard way to create an ABC using
HotDistanceLoss	Helper class that provides a standard way to create an ABC using

class dacapo.experiments.tasks.losses.DummyLoss

A class representing a dummy loss function that calculates the absolute difference between each prediction and target.

Inherits the Loss class.

compute(prediction, target, weight=None)

Calculate the total loss between prediction and target.

compute(prediction, target, weight=None)

Method to calculate the total dummy loss.

Parameters:

• prediction (float or int) – predicted output

• target (float or int) – true output

• weight (float or int, optional) – weight parameter for the loss, by default None

Returns:

Total loss calculated as the sum of absolute differences between prediction and target.

Return type:

float or int

class dacapo.experiments.tasks.losses.MSELoss

A class used to represent the Mean Square Error Loss function (MSELoss).

None

compute(prediction, target, weight):

Computes the MSELoss with the given weight for the predictiom and target.

compute(prediction, target, weight)

Function to compute the MSELoss for the provided prediction and target, with respect to the weight.

Parameters:

predictiontorch.Tensor

The prediction tensor for which loss needs to be calculated.

targettorch.Tensor

The target tensor with respect to which loss is calculated.

weighttorch.Tensor

The weight tensor used to weigh the prediction in the loss calculation.

Returns:

: torch.Tensor

The computed MSELoss tensor.

class dacapo.experiments.tasks.losses.Loss

Helper class that provides a standard way to create an ABC using inheritance.

abstract compute(prediction: torch.Tensor, target: torch.Tensor, weight: torch.Tensor | None = None) → torch.Tensor

Compute the loss for the given prediction and target. Optionally, if given, a loss weight should be considered.

All arguments are torch tensors. The return type should be a torch scalar that can be used with an optimizer, just as usual when training with torch.

class dacapo.experiments.tasks.losses.AffinitiesLoss(num_affinities: int, lsds_to_affs_weight_ratio: float)

Helper class that provides a standard way to create an ABC using inheritance.

compute(prediction, target, weight)

Compute the loss for the given prediction and target. Optionally, if given, a loss weight should be considered.

All arguments are torch tensors. The return type should be a torch scalar that can be used with an optimizer, just as usual when training with torch.

class dacapo.experiments.tasks.losses.HotDistanceLoss

Helper class that provides a standard way to create an ABC using inheritance.

compute(prediction, target, weight)

Compute the loss for the given prediction and target. Optionally, if given, a loss weight should be considered.

All arguments are torch tensors. The return type should be a torch scalar that can be used with an optimizer, just as usual when training with torch.

hot_loss(prediction, target, weight)

distance_loss(prediction, target, weight)

split(x)