Module PINNacle.src.optimizer.ntk

Classes

LR_Adaptor_NTK(optimizer, loss_weight, pde) : Callback for learning rate annealing algorithm of physics-informed neural networks.

loss_weight - initial loss weights
num_pde - the number of the PDEs (boundary conditions excluded)

### Ancestors (in MRO)

* torch.optim.optimizer.Optimizer

### Methods

`step(self, closure)`
:

Multiadam PDE