fix: add graphsage activation + torch preprocess for trinet vision#31
fix: add graphsage activation + torch preprocess for trinet vision#31melvinbarbaux merged 3 commits intomainfrom
Conversation
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: c866230697
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| if hidden_sizes: | ||
| resolved_hidden = int(hidden_sizes[0]) | ||
| if resolved_layers is None: | ||
| resolved_layers = max(2, int(len(hidden_sizes)) + 1) |
There was a problem hiding this comment.
Respect all hidden_sizes entries when building GraphSAGE
When hidden_sizes has more than one entry (e.g., [128, 64]), the code only reads hidden_sizes[0] but still bumps num_layers based on the full length. This means the additional sizes are silently ignored and all hidden layers use the first size, so the model architecture does not match the user-configured hyperparameters and can change experiment outcomes. Consider constructing per-layer sizes from the full list or rejecting lists longer than 1.
Useful? React with 👍 / 👎.
b048448 to
ada6d79
Compare
Codecov Report✅ All modified and coverable lines are covered by tests. 📢 Thoughts on this report? Let us know! |
Co-authored-by: Copilot Autofix powered by AI <223894421+github-code-quality[bot]@users.noreply.github.com>
Summary
What does this PR change?
Checklist
Notes
Anything reviewers should know?