kaira.losses.audio.FeatureMatchingLoss

Inheritance diagram for FeatureMatchingLoss

class kaira.losses.audio.FeatureMatchingLoss(model, layers, weights=None)

Bases: BaseLoss

Feature Matching Loss Module.

This module calculates the loss between features extracted from a pretrained model.

Methods

__init__	Initialize the FeatureMatchingLoss module.
forward	Forward pass through the FeatureMatchingLoss module.

__init__(model, layers, weights=None)

Initialize the FeatureMatchingLoss module.

Parameters:

• model (BaseLoss) – Pretrained model for feature extraction.

• layers (list) – List of layer indices to extract features from.

• weights (list, optional) – Weights for each layer. Default is None (equal weights).

forward(x: Tensor, target: Tensor) → Tensor

Forward pass through the FeatureMatchingLoss module.

Parameters:

• x (torch.Tensor) – The input audio tensor.

• target (torch.Tensor) – The target audio tensor.

Returns:

The feature matching loss.

Return type:

torch.Tensor