deepmd.pt.loss.denoise

Classes

DenoiseLoss

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

Module Contents

class deepmd.pt.loss.denoise.DenoiseLoss(ntypes: int, masked_token_loss: float = 1.0, masked_coord_loss: float = 1.0, norm_loss: float = 0.01, use_l1: bool = True, beta: float = 1.0, mask_loss_coord: bool = True, mask_loss_token: bool = True, **kwargs: Any)

Bases: deepmd.pt.loss.loss.TaskLoss

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

ntypes

masked_token_loss = 1.0

masked_coord_loss = 1.0

norm_loss = 0.01

has_coord

has_token

has_norm

use_l1 = True

beta = 1.0

frac_beta = 1.0

mask_loss_coord = True

mask_loss_token = True

forward(model_pred: dict[str, torch.Tensor], label: dict[str, torch.Tensor], natoms: int, learning_rate: float, mae: bool = False) → tuple[torch.Tensor, dict[str, torch.Tensor]]

Return loss on coord and type denoise.

Returns:

• loss: Loss to minimize.