Open 4fee8fea opened 2 years ago
Thanks for the question. This is a good feature request. Currently, I recommend trying out the design space of 96 GNN models, shown here: https://github.com/snap-stanford/GraphGym/blob/master/run/grids/example.txt
You could take a look at the jupyter notebook here, regarding how to compute task similarity: https://github.com/snap-stanford/GraphGym/blob/master/analysis/design_space.ipynb
We are also actively working on enriching the functionalities of GraphGym.
Thanks for the exciting program!
I am suffering from finding the optimal GNN model on the node classification task. This problem is caused by too much freedom of choice within and between layers. In other words, there are too many models to choose from and too many hyperparameters to optimize.
Referring to ogb's leaderboard to find the optimal model is a potential solution, but as the paper showed,
From my understanding, GraphGym has provided the idea that similar tasks can share the optimal model design.
As mentioned,
I would like to know if there are off-the-shelf model-task combinations that I can use directly, without using the interface to try out GNN designs.