cszhangzhen / mvpool Goto Github PK

Hi, this paper seems good, but I cannot reproduce the reported performance based on given code.
Especially on some node classification datasets, the difference between the results we reproduce and those given in the paper is huge.
Is there anything that needs extra attention? For example, the number of pooled layers, the pooled ratio of each layer, or other important settings.

Hello, how to solve this problem?

File "/home/adminm/gm2/MVPool/sparse_softmax.py", line 29, in scatter_sort
    batch_size, max_num_nodes = num_nodes.size(0), num_nodes.max().item()
RuntimeError: max(): Expected reduction dim to be specified for input.numel() == 0. Specify the reduction dim with the 'dim' argument.