drugforge.ml.loss.RangeLoss

class drugforge.ml.loss.RangeLoss(*args: Any, **kwargs: Any)

Bases: Module

__init__(lower_lim, upper_lim)

Class for calculating a loss to penalize predictions outside of the given range. Current implementation uses a squared difference penalty.

Parameters:

• lower_lim (float) – Bottom limit of acceptable range

• upper_lim (float) – Upper limit of acceptable range

Methods

__init__(lower_lim, upper_lim)	Class for calculating a loss to penalize predictions outside of the given range.
forward(pred, pose_preds, target, in_range, ...)	No loss for predictions within self range, otherwise calculate squared distance to closest bound.

forward(pred, pose_preds, target, in_range, uncertainty)

No loss for predictions within self range, otherwise calculate squared distance to closest bound.

Parameters:

• pred (torch.Tensor) – Model prediction

• target (torch.Tensor) – Prediction target

• in_range (torch.Tensor) – target’s presence in the dynamic range of the assay. Give a value of < 0 for target below lower bound, > 0 for target above upper bound, and 0 or None for inside range

• uncertainty (torch.Tensor) – Uncertainty in target measurements

Returns:

Calculated loss

Return type:

torch.Tensor