GConvGRU hidden state not passed

benedekrozemberczki / pytorch_geometric_temporal

PyTorch Geometric Temporal: Spatiotemporal Signal Processing with Neural Machine Learning Models (CIKM 2021)

MIT License

2.62k stars 369 forks source link

GConvGRU hidden state not passed #145

Closed josephenguehard closed 2 years ago

josephenguehard commented 2 years ago

Hi! I have a similar question as #143: in the GConvGRU examples recurrent folder, the hidden state is not passed at all. Would that not harm performance as the hidden state is then always reset to zeros in the for loop?

benedekrozemberczki commented 2 years ago

Yes it would, you can set the cell states between the epoch and the sequence iteration loop.

josephenguehard commented 2 years ago

Thank you for your quick reply!