topobench.loss.model package

Submodules

topobench.loss.model.DGMLoss module

Differentiable Graph Module loss function.

class topobench.loss.model.DGMLoss.DGMLoss(loss_weight=0.5)

Bases: AbstractLoss

DGM loss function.

Original implementation lcosmo/DGM_pytorch

Parameters:

loss_weightfloat, optional

Loss weight (default: 0.5).

forward(model_out: dict, batch: Data) → Tensor

Forward pass of the loss function.

Parameters:

model_outdict

Dictionary containing the model output.

batchtorch_geometric.data.Data

Batch object containing the batched domain data.

Returns:

dict

Dictionary containing the model output with the loss.

topobench.loss.model.GraphMLPLoss module

Graph MLP loss function.

class topobench.loss.model.GraphMLPLoss.GraphMLPLoss(r_adj_power=2, tau=1.0, loss_weight=0.5)

Bases: AbstractLoss

Graph MLP loss function.

Parameters:

r_adj_powerint, optional

Power of the adjacency matrix (default: 2).

taufloat, optional

Temperature parameter (default: 1).

loss_weightfloat, optional

Loss weight (default: 0.5).

forward(model_out: dict, batch: Data) → Tensor

Forward pass of the loss function.

Parameters:

model_outdict

Dictionary containing the model output.

batchtorch_geometric.data.Data

Batch object containing the batched domain data.

Returns:

dict

Dictionary containing the model output with the loss.

get_power_adj(edge_index)

Get the power of the adjacency matrix.

Parameters:

edge_indextorch.Tensor

Edge index tensor.

Returns:

torch.Tensor

Power of the adjacency matrix.

graph_mlp_contrast_loss(x_dis, adj_label)

Graph MLP contrastive loss.

Parameters:

x_distorch.Tensor

Distance matrix.

adj_labeltorch.Tensor

Adjacency matrix.

Returns:

torch.Tensor

Contrastive loss.

Module contents

This module implements the loss functions for the topobench package.