Reformat the CNODE structure #47
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
We now use view+reshape to allow the manipulation of the same data in two ways * linear view, in order to easily concatenate fields with different grid sizes in an array of size (linear, batch) * grid view, in order to implement discretized forces in a more natural way In the current implementation, there is almost no overhead in using either of the two shapes. Currently the only example that has been adapted is 02.01.
Remove reshape
Improve Laplacian
Use only grid view via ArrayPartition
* Move things around: - Delete test.jl didn't have anything meaningful. - Delete INS_SciMLInterface.jl: code there was outdated. We've developed a better version of it in INS/lib/SciMLCompat. - Create subfolders in src and adapt CoupleNODE.jl to expose the memebers accordingly. - Delete circular import of grid. - Delete grid.jl. Most probably we are going to use INS, but we can always check this file back in the tags. - Update NS_solvers.jl to use the function in CoupleNODE (similar to the one in lib/SciMLCompat). - Update NS_closure.jl to match the zygote part of scripts Enzyve_vs_Zygote from SciMLCompat. * Proposes the workflow for generating data and training via apriori and a-posteriori fitting in the script simulations/NavierStokes_2D/scripts/NeuralClosure+SciML.jl. We've also added a module NavierStokes that contains utilities necessary for this purpose. --------- Co-authored-by: Luisa Orozco <99738896+luisaforozco@users.noreply.github.com> Co-authored-by: Victor Azizi <victor@lipsum.eu> Co-authored-by: SCiarella <58949181+SCiarella@users.noreply.github.com>
Cleanup: create folder helper_scripts
|
LGTM, |
…e the data. Now, you can load the data via JLD2 and contnue from there.
Remove NeuralClosure dependency
Fix aposteriori
This was referenced Oct 2, 2024
Closed
Closed
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This branch removes a lot of overhead from the
CNODEstruct.This is a major change from the main branch, but the performance gains justify it completely.
Basically this branch shows that pre and pos force operations are not needed.
At the moment, only the Burgers examples have been updated.