Alvaro
Alvaro
I would refer to the main paper for question 1. We trained for 20M steps, but usually performance at 2.5M steps is already very good. Training this model on a...
Thank you unfortunately we will not be able to share the train model weights at this point, but I would encourage you to find other GitHub issues where other users...
Unfortunately this implementation is provided for a simple illustrative reference of the model to go with the paper and to show some basic training, and is not really meant to...
Thanks for your message! My interpretation is that catastrophic forgetting is not an issue, because even though it is true that the number of effective training examples is in the...
Do other regular matplotlib animations work for you (for example some animations from [here](https://matplotlib.org/stable/api/animation_api.html)), or is this specific to our own code? If the matplotlib examples have the same issue,...
Thanks for the message, unfortunately this code provided as is, but documenting dataset generation is beyond the scope of this reference implementation of the model. That said, you may find...
> I can't make sure my understanding of your dataset is true Thanks for your message, yes, that description seems accurate. > The data of the velocity and pressure is...
> I've tried adding global variables to each node. But it doesn't work well. All we do in this paper is to concatenate the global feature to all of the...