arXiv link?

[scaomath]

Hi @Till-M , cool repo. I like that you included a standalone architecture file for us to try. Can you post the paper name and arXiv link as well?

So adding a learnable displacement in each level is the secret sauce for fluid? That is pretty neat.

https://github.com/till-m/flowers/blob/8db4187e3eb38f448faaaa741370aa406d56b0c5/flower_standalone.py#L534C5-L536C44

[t-muser]

Hey @scaomath,

thanks for your interest! I submitted the paper to arxiv a week ago and it has been "on hold" since then. I will update the repo with an arxiv link as soon as it gets released. For now, you can read the paper on researchgate: https://www.researchgate.net/publication/400979038_Flowers_A_Warp_Drive_for_Neural_PDE_Solvers

That is pretty neat.

Indeed! :)

[t-muser]

Hey, paper is now out on arxiv: https://arxiv.org/abs/2603.04430, closing this accordingly.

PS: I figured out why your name rang a bell. Nice paper on the Galerkin transformer ;)

[Eydcao]

Nice work!

Just realize this is the official repo for the arxiv (yeah I came from another way around).

Can you update the arxiv to include GitHub link too?

Best

[t-muser]

Hey @Eydcao,

Can you update the arxiv to include GitHub link too?

We're planning to update the paper anyways, I will add this to the list of things! It will probably take a few weeks though :) I will re-open the issue as a reminder.

Nice work!

Thanks! If you end up using the model for some task please let me know! :)