Open jwallwork23 opened 1 year ago
I think subclassing GoalOrientedMeshSeq would be a nice way to train a neural network to do goal-oriented error estimation.
GoalOrientedMeshSeq
Sketch implementation here: https://github.com/pyroteus/pyroteus/tree/121_dd_mesh_seq.
I think subclassing
GoalOrientedMeshSeq
would be a nice way to train a neural network to do goal-oriented error estimation.