Open kingwmk opened 5 years ago
Currently, this is not supported, unfortunately. The best way is probably to train on shorter sequences if you expect the latent graph to change quickly. Evaluation with dynamic_graph=True should then improve.
On Mon, Jan 14, 2019 at 10:27 AM kingwmk notifications@github.com wrote:
Hi, thanks for your outstanding works and contribution. I have a question that Can we use dynamic_graph in training step? If yes, can you give me some implementation guidance? Thank you very much!
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/ethanfetaya/NRI/issues/10, or mute the thread https://github.com/notifications/unsubscribe-auth/AHAcYLH0iI0DeB6X8wBUyCWXdGAsPHh3ks5vDE1rgaJpZM4Z93wa .
Thank you !!!
Hi, thanks for your outstanding works and contribution. I have a question that Can we use dynamic_graph in training step? If yes, can you give me some implementation guidance? Thank you very much!