TopoModelX
TopoModelX copied to clipboard
pyt-team
Topological Deep Learning
Two important concerns about `SCConv` simplicial model: - [ ] **Implementation of normalization**: The original paper consider normalized incidence and adjacency matrices, but the proposed normalization procedure in the tutorial...
Review the model implementations and make sure they converge properly on real-world datasets. nn/cell: - [x] CAN - [x] CCXN - [x] CWN nn/combinatorial - [x] HMC
Add a separate numpydoc validation step to the projects CI setup, similar to pyt-team/TopoNetX#318.
# What? Create a [pyt-team](https://pyt-team.github.io/home/) website that serves as a single entry point that points towards the three documentation websites. # Why? Users might want to circulate from one doc...
# What? This is the list of essentials that need to be checked before submitting the software paper. - [x] Tests are passing on all repositories. - [x] Code coverage...
# What? Some docstrings are erroneous, for example: ``` class SCNN(torch.nn.Module): """Simplicial convolutional neural network implementation for complex classification. Note: At the last layer, we obtain the output on simplcies,...
While `Dist2CycleLayer`s are [correctly constructed](https://github.com/pyt-team/TopoModelX/blob/e4a893fcb2e99c9868b1423c41bb4d90893f85bc/topomodelx/nn/simplicial/dist2cycle.py#L19) in `Dist2Cyle`, they are [never actually called](https://github.com/pyt-team/TopoModelX/blob/e4a893fcb2e99c9868b1423c41bb4d90893f85bc/topomodelx/nn/simplicial/dist2cycle.py#L55). - [ ] Correctly call layers in the `forward` method. - [ ] This should have been catched...