You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
@kayzliu
When I write the Dominant example, I find the following issues. Please fix/answer them accordingly.
The current process_graph function is dedicated to the BlogCatalog dataset, we need to write a general dataloader that could handle any PyG data object. The preprocessing code for BlogCatalog can be put into the dominant.py under /example.
When I run model.fit(), train_loss became NaN after 5-6 epochs.
How is the outlier label of BlogCatalog generated?
Should we train the model on clean data and evaluate it on data with outliers?
The text was updated successfully, but these errors were encountered:
The BlogCatalog dataset is for code validation only. The labels make no sense. I have updated the code in the latest commit. It should work well with correct outlier labels. Shall we override the original labels in the dataset with outlier labels (in
pygod/utils/outlier_generator.py)?
@kayzliu
When I write the Dominant example, I find the following issues. Please fix/answer them accordingly.
process_graph
function is dedicated to the BlogCatalog dataset, we need to write a general dataloader that could handle any PyG data object. The preprocessing code for BlogCatalog can be put into thedominant.py
under/example
.model.fit()
, train_loss becameNaN
after 5-6 epochs.The text was updated successfully, but these errors were encountered: