Hi~ I'm very interested in your paper, and I want to run your code.
I found that in the BaseTask.py you set batch_size=1, it seems that it should be used for blockchain graph.
I think larger batch size should accelerate in other database like QM9. But when batch size is large than 1, the error occurs. It seems that due to you have not write code to pad graphs with different size in a batch. Could you tell me is it possible to run with larger batch size?