Dear Xiaolan Xie,

thank you for your interest in our work. The dimension 32 from the link_inputs comes from the link hidden state size. The other two dimensions are not fixed, that's why it appears with the symbol "?". This is because there can be different number of paths and links for each topology. Therefore, with "?" it indicates that the dimension is variable. For more details about the shape of link_inputs see https://github.com/knowledgedefinednetworking/demo-routenet/blob/6fe560081f8....

Best,

Paul Almasan

On Mon, 20 Feb 2023 at 07:51, 谢小兰 21032303148@mails.guet.edu.cn wrote:

Dear Pro J. Suárez-Varela, S. Carol-Bosch, K. Rusek, P. Almasan, M. Arias, P. Barlet-Ros, A. Cabellos-Aparicio:

Hello Pro J. Suárez-Varela, S. Carol-Bosch, K. Rusek, P. Almasan, M. Arias, P. Barlet-Ros, A. Cabellos-Aparicio, I am sorry to disturb you in your busy schedule. I'am a student of Guilin University of Electronic Science and Technology from China.

Recently, Iam learning how to do route optimization with grahp neural networks. Your paper RouteNet: Leveraging Graph Neural Networks for network modeling and optimization in SDN gives me great inspiration. I ran the model code after reading your paper, but I can't see how much about the model regarding the input and output dimensions, for example, the shape of link_inputs=(? ,? ,32). I would like to ask about this? What is the size of the How can I can see it in the code?

No matter whether you help me or not, best wishes to you. I would appreciate it if you could help me.

Thanks and regards. Sincerely, Xiaolan, Xie 2023/2/19

---

Kdn-users mailing list Kdn-users@knowledgedefinednetworking.org https://mail.n3cat.upc.edu/cgi-bin/mailman/listinfo/kdn-users