nntoolbox.vision.losses.style module

class nntoolbox.vision.losses.style.FeatureLoss(model, layers, base_loss=<class 'torch.nn.modules.loss.MSELoss'>)

Bases: torch.nn.modules.module.Module

compute_features(output, target)

forward(output, target)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

training: bool

class nntoolbox.vision.losses.style.INStatisticsMatchingStyleLoss(model, layers, base_loss=<class 'torch.nn.modules.loss.MSELoss'>)

Bases: nntoolbox.vision.losses.style.FeatureLoss

As suggested by https://arxiv.org/pdf/1703.06868.pdf

compute_features(output, target)

training: bool

class nntoolbox.vision.losses.style.StyleLoss(model, layers, base_loss=<class 'torch.nn.modules.loss.MSELoss'>)

Bases: nntoolbox.vision.losses.style.FeatureLoss

compute_features(output, target)

gram_mat(features)

training: bool

class nntoolbox.vision.losses.style.TotalVariationLoss(base_loss=<class 'torch.nn.modules.loss.L1Loss'>)

Bases: torch.nn.modules.module.Module

Based on https://github.com/tensorflow/tensorflow/blob/r1.13/tensorflow/python/ops/image_ops_impl.py

forward(input: torch.Tensor) → torch.Tensor

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

training: bool