search

Spider-scnu / TSP

MIT License
111 stars 18 forks source link

More on Att-GCRN #2

Open chaitjo opened 3 years ago

chaitjo commented 3 years ago

Hi @Spider-scnu, thank you for the super insightful work and for releasing the code!

I had a quick question re. the GNN model used 'Att-GCRN'. Is this model any different from the model proposed in Joshi-etal, 2019? On looking at the code, it seems the same overall architecture.

Best, Chaitanya

chaitjo commented 3 years ago

Hi @Spider-scnu, just wanted to ping you on this once again.

Spider-scnu commented 2 years ago

Hi, the code of backbone Att-GCRN is baed on Joshi-etal, 2019, and also I have add training tricks and done some modification about the implementation of attention. In our paper, we stressed the importance of sampling techniques and our MCTS improvement methods to solve the big TSP instances.

Best, Kaibin QIU

At 2021-03-09 13:46:47, "Chaitanya Joshi" @.***> wrote:

Hi @Spider-scnu, thank you for the super insightful work and for releasing the code!

I had a quick question re. the GNN model used 'Att-GCRN'. Is this model any different from the model proposed in Joshi-etal, 2019? On looking at the code, it seems the same overall architecture.

Best, Chaitanya

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.