Given input size: (1x9x1024). Calculated output size: (1x-1x1024). Output size is too smal

[XuMengyaAmy]

Thanks for your great work. I am reproducing the results from the paper "Graph convolutional nets for tool presence detection in surgical videos".

I face the error. "RuntimeError: Given input size: (1x9x1024). Calculated output size: (1x-1x1024). Output size is too small"
It is related with the pool layer in the model architecture in https://github.com/uta-smile/STGCN-IPMI19/blob/main/models/video_models.py
Can I check your input shape?
output = model(data, adj) data.shape is torch.Size([64, 9, 1024]) and adj.shape is torch.Size([64, 9, 9])

Can you help me to solve this issue?
Thanks.