Open anubane opened 2 months ago
While executing the ranking logic with GNNExplainer, it seems that the function predict_candidate_genes_gnn_explainer_only in file GDARanking.py has a memory leak issue!
predict_candidate_genes_gnn_explainer_only
GDARanking.py
When the candidates population logic is being run, the memory consumed keeps growing.
candidates
I think that this is caused by run-away creation of numerous small tensor objects in the candidates dictionary.
It would seem that the same issue may exist with the other ranker functions as well!
Thank you for reaching out. We will check the problem and, in case, provide a fix with the next code update. Stay tuned 🧑💻
While executing the ranking logic with GNNExplainer, it seems that the function
predict_candidate_genes_gnn_explainer_only
in fileGDARanking.py
has a memory leak issue!When the
candidates
population logic is being run, the memory consumed keeps growing.I think that this is caused by run-away creation of numerous small tensor objects in the
candidates
dictionary.