David Pitt
David Pitt
Hi @mshmoon, thank you for this feedback. This is really good to know from a user! Since last month we've made some updates to the GINO model including the docstring...
@mshmoon - you can check `scripts/train_gino_carcfd.py` for an example!
Hi @profPlum, you're right that the base FNO isn't designed to handle irregular meshes. However, there are models like the one in the paper you linked (for instance, one that...
Hi @profPlum , in our latest round of updates we now expose the `GINO` model to map between irregular meshes.
> Wouldn't the default **deepcopy** work out of the box? No, it throws an error - there's no way to access the internal module state from base_model by default.
As of #472, this issue is resolved. @gokhalen , please refer to `neuralop.data.datasets`. Thanks!
Hi @rochaie, is this issue related to the `neuralop` library? If not this might be better suited for an issue in the repo which contains the source mentioned here
Closing because this is stale and doesn't refer to the current library code
regularizer should be a PyTorch regualizer object
Currently fixed in `trainer`: ``` if self.regularizer: self.regularizer.reset() ```