Benchmark about neuraloperator HOT 13 CLOSED

Figure 2 in the paper seems to be the architecture. Do you mean the experiments on Burgers, Darcy, or Navier-Stokes?

from neuraloperator.

Oh I see, I see. We haven't decided to publish all the baseline methods, since they need a lot of clean-up. But if you don't mind, I can send you some of these baselines we have at hand. If that can be helpful, please let me know your email address. Thanks.

from neuraloperator.

While I am preparing the code, please see these repositories:

• For GKN and MKGN: https://github.com/zongyi-li/graph-pde
• For UNets, Resnets, TF-Nets: https://github.com/Rose-STL-Lab/Turbulent-Flow-Net
• The lowrank method is contained in this repostory (e.g. lowrank_2d.py)

I when send you the code for FCN, PCANN, and ROM once I gather them. Thanks!

from neuraloperator.

Sorry, I had an earlier version. Yes, figure 3 - Benchmark on Burger's equation, Darcy Flow, and Navier-Stokes equation

from neuraloperator.

Sorry, do you mean you cannot reproduce the results of FNO on any of these datasets? or do you want to test with the baseline methods?

from neuraloperator.

I already reproduce for the FNO! I would like to reproduce for the other methods, such as NN, FCN, PCAMM, RBM and LNO. Do you have this code available, for example for the 2-d Darcy Flow equation?

My goal is to obtain the results from table 4 for the different methods. Thanks

from neuraloperator.

I understand perfectly, so if you don't mind, I would ask you to send it. I'm sure that it will help me. Very grateful for your help, Zongyi. My email is: [email protected]. Thanks!

from neuraloperator.

Thank you very much! I am already analyzing the repositories you sent.
Then when you get the code for FCN, PCANN, and RBM, please send. Thanks, Zongyi!

from neuraloperator.

Hi Zongyi!

I am trying to reproduce the results right now. If you don't mind, is it possible for you to send me a copy of the code for the baseline results as well (if you have it)? That would be really helpful!

My email address is [email protected].

Thanks a lot!

Best,
Kai

from neuraloperator.

Hi @zongyi-li I'm also interested in the baseline model, especially U-Net implementation. Did you use it autoregressively in a closed loop manner? It would be great if you could also send me your implementation. My email address: [email protected]
Many thanks in advance! :D

from neuraloperator.

Hi Zongyi!
I am trying to reproduce the inverse problem right now. If you don't mind, is it possible for you to send me a copy of the code for the inverse probelm. That would be really helpful!
My email address is [email protected]
Thanks a lot!

Best,
Ming

from neuraloperator.

Same here, would appreciate a copy of the code to replicate the benchmarks to [email protected]. Thanks!

from neuraloperator.

Closing as inactive. Feel free to reopen.

from neuraloperator.