TGAT MRR - #58
Merged
Merged
Conversation
Jacob-Chmura
marked this pull request as draft
June 24, 2025 17:07
shenyangHuang
approved these changes
Jun 30, 2025
| device = batch.src.device | ||
| z = torch.zeros(len(batch.unique_nids), self.embed_dim, device=device) | ||
|
|
||
| for hop in reversed(range(self.num_layers)): |
Collaborator
There was a problem hiding this comment.
maybe let's walk through this together when we meet
Member
Author
There was a problem hiding this comment.
Sounds good. This is just how message passing is implemented. You always have to start at your boundary (highest hop)
| # we pick random time stamps within temporal window of the batch. | ||
| # Using random times on the whole graph will likely produce information | ||
| # leakage, making the prediction easier than it should be. | ||
| fake_times = torch.randint( |
Collaborator
There was a problem hiding this comment.
yeah I think this is fine for now. In the ideal case, you should produce timestamp for fake edges based on the true edge that it corresponds to, this is a sufficient condition for us now. We can also address this later on
Codecov ReportAttention: Patch coverage is
📢 Thoughts on this report? Let us know! |
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Purpose
The purpose of this PR is trying to replicate and patch #57. I extracted the TGAT mrr based example to try and isolate the problem.
Performance
Note: our validation MRR after a single epoch is 0.0740. The TGB leaderboard expects 0.131 ± 0.008 (after 10 epochs). A randomly initialized TGAT model (garbage weights) achieves MRR = 0.0438, so something is happening.
Probably training longer would help. As would using actual static node features (or gaussian random) instead of zeros. There are probably also actual bugs.
Key Changes
Relevant Prs
Close #57
Out of scope