d-ailin/GDN

Error in training

Closed this issue · 3 comments

I'm trying to use your implementation for a project, while running the train function I get this error: RuntimeError: mat1 and mat2 shapes cannot be multiplied (1x55 and 15x64).

55 is the number of features of my dataset
15 sliding window
64 dim

Thanks for your interest. Could you provide more details, such as the error stack trace and the config?

I guess it is caused by the incorrect input data format. Here, the input's shape should be in something like (batch_size, slide_win).

The input's shape is actually torch.Size([batch_size, n_nodes, slide_win]).
Is it correct? I get very high loss value, so I think I'm making some mistakes..