Hypergraph checks #249
Merged
Hypergraph checks #249
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
added 14 commits
October 30, 2023 15:36
levtelyatnikov
requested review from
ninamiolane,
ManuelLecha,
mathildepapillon and
mhajij
11 tasks
|
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
ninamiolane
requested changes
Nov 6, 2023
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #249 +/- ##
==========================================
- Coverage 97.49% 97.45% -0.05%
==========================================
Files 58 57 -1
Lines 2155 2121 -34
==========================================
- Hits 2101 2067 -34
Misses 54 54
☔ View full report in Codecov by Sentry. |
ninamiolane
approved these changes
Nov 7, 2023
topomodelx/nn/hypergraph/allset.py
Outdated
| return torch.sigmoid(self.linear(pooled_x))[0] | ||
| x_0, x_1 = layer(x_0, incidence_1) | ||
|
|
||
| return (x_0, x_1) |
| mlp_num_layers : int, default: 2 | ||
| Number of layers in the MLP. | ||
| mlp_activation : torch.nn.Module, default: None | ||
| Activation function for the MLP. | ||
| mlp_dropout : float, default: 0.0 | ||
| Dropout probability for the MLP. |
There was a problem hiding this comment.
missing parameter type: e.g. dropout : float and note the space before : according to numpy's conventions for docstring, which we use here.
topomodelx/nn/hypergraph/allset.py
Outdated
| mlp_dropout : float, default: 0.0 | ||
| Dropout probability for the MLP. | ||
| mlp_activation: |
|
|
||
| return x | ||
| return (x_0, x_1) |
| Number of AllSet layers in the network. | ||
| heads : int, default: 4 | ||
| Number of attention heads. | ||
| dropout : float, default=0.2 | ||
| dropout : float |
topomodelx/nn/hypergraph/unigin.py
Outdated
| self.input_drop = torch.nn.Dropout(input_drop) | ||
| self.layer_drop = torch.nn.Dropout(layer_drop) | ||
|
|
||
| # Define initial linear layer |
topomodelx/nn/hypergraph/unigin.py
Outdated
| x_0 = self.layer_drop(x_0) | ||
| x_0 = torch.nn.functional.relu(x_0) | ||
|
|
||
| return (x_0, x_1) |
| # Update node features using GIN update equation | ||
| return self.nn((1 + self.eps) * x_0 + m_1_0) | ||
| x_0 = self.linear((1 + self.eps) * x_0 + m_1_0) | ||
| return (x_0, x_1) |
topomodelx/nn/hypergraph/unisage.py
Outdated
| x_0 = self.layer_drop(x_0) | ||
| x_0 = torch.nn.functional.relu(x_0) | ||
|
|
||
| return (x_0, x_1) |
| m_1_0 = self.edge2vertex(x_1, incidence_1) | ||
| x_0 = x_0 + m_1_0 | ||
|
|
||
| return (x_0, x_1) |
ninamiolane
approved these changes
Nov 9, 2023
This was referenced Feb 14, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Overall:
I have reviewed the hypergraph implementation and made necessary modifications to ensure consistent patterns. Additionally, I have updated tutorials and test files to align with these changes.
Specifically:
Previously, all the models generated the 'desired' hidden representation, meaning that if a model was required to perform a graph classification task, the pooling operation was executed within the model. Hence, the model output was the pooled signal.
Now, the models output the last hidden representations. In hypergraph models, the output includes nodes and hyperedges.
We have introduced a separate, lightweight pooling/readout class that is responsible for providing the desired output for particular downstream tasks. This new pooling/readout class is currently introduced in the tutorials.