l2gv2.align.geo package

l2gv2.align.geo package#

Submodules#

l2gv2.align.geo.geoalign module#

Alignment using pytorch geometric.

class l2gv2.align.geo.geoalign.GeoAlignmentProblem(patches: list[Patch], patch_edges=None, min_overlap=None, copy_data=True, self_loops=False, verbose=False, num_epochs: int = 1000, learning_rate: float = 0.001, model_type: str = 'affine', device: str = 'cpu')#

Bases: AlignmentProblem

Alignment problem using pytorch geometric.

align_patches(scale=False)#

Align the patches.

l2gv2.align.geo.model module#

Model for aligning patch embeddings

class l2gv2.align.geo.model.AffineModel(dim, n_patches, device)#

Bases: Module

Model for aligning patch embeddings

forward(patch_intersection)#

Forward pass

class l2gv2.align.geo.model.OrthogonalModel(dim, n_patches, device)#

Bases: Module

Model for aligning patch embeddings

forward(patch_intersection)#

Forward pass :param patch_intersection: list of tuples

Returns:

list of transformed embeddings

l2gv2.align.geo.train module#

Train the model on the patch embeddings

l2gv2.align.geo.train.patchgraph_mse_loss(transformed_emb)#

Custom loss function that computes the squared norm of differences between transformed pairs in the dictionary.

Parameters:

transformed_dict – Dictionary with keys (i,j) and values (XW_i+b_i, YW_j+b_j)

Returns:

Total loss as the sum of squared differences

l2gv2.align.geo.train.train_alignment_model(patch_intersections, n_patches, device='cpu', num_epochs=100, learning_rate=0.05, model_type='affine', verbose=True)#

Train the model on the patch embeddings :param patch_emb: list of torch.Tensor

patch embeddings

Parameters:

  • n_patches – int number of patches

  • device – str device to run the model on

  • num_epochs – int number of epochs to train the model

  • learning_rate – float learning rate for the optimizer

Returns:

Model loss_hist: list

Return type:

model

Module contents#

class l2gv2.align.geo.GeoAlignmentProblem(patches: list[Patch], patch_edges=None, min_overlap=None, copy_data=True, self_loops=False, verbose=False, num_epochs: int = 1000, learning_rate: float = 0.001, model_type: str = 'affine', device: str = 'cpu')#

Bases: AlignmentProblem

Alignment problem using pytorch geometric.

align_patches(scale=False)#

Align the patches.

class l2gv2.align.geo.OrthogonalModel(dim, n_patches, device)#

Bases: Module

Model for aligning patch embeddings

forward(patch_intersection)#

Forward pass :param patch_intersection: list of tuples

Returns:

list of transformed embeddings

l2gv2.align.geo.patchgraph_mse_loss(transformed_emb)#

Custom loss function that computes the squared norm of differences between transformed pairs in the dictionary.

Parameters:

transformed_dict – Dictionary with keys (i,j) and values (XW_i+b_i, YW_j+b_j)

Returns:

Total loss as the sum of squared differences

l2gv2.align.geo.train_alignment_model(patch_intersections, n_patches, device='cpu', num_epochs=100, learning_rate=0.05, model_type='affine', verbose=True)#

Train the model on the patch embeddings :param patch_emb: list of torch.Tensor

patch embeddings

Parameters:

  • n_patches – int number of patches

  • device – str device to run the model on

  • num_epochs – int number of epochs to train the model

  • learning_rate – float learning rate for the optimizer

Returns:

Model loss_hist: list

Return type:

model

Contents