You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Is your feature request related to a problem? Please describe.
From PyTorch 2.0 there is the possibility to compile the code, resulting in faster training and inference time, see here for reference. Also PyTorch Lightning supports the torch compilation as shown here.
It would be nice to implement a compilation flag to all PINA solvers to enable compilation for the model and equations (if present in the problem).
Describe the solution you'd like
This could be implemented by using the two abstract class SolverInterface, Equation. The amount of code change should be minimal, since PyTorch Lightning already deals itself to compile all internal functions (training_step, forward, ...).
The text was updated successfully, but these errors were encountered:
Is your feature request related to a problem? Please describe.
From PyTorch 2.0 there is the possibility to compile the code, resulting in faster training and inference time, see here for reference. Also PyTorch Lightning supports the torch compilation as shown here.
It would be nice to implement a compilation flag to all PINA solvers to enable compilation for the model and equations (if present in the problem).
Describe the solution you'd like
This could be implemented by using the two abstract class SolverInterface, Equation. The amount of code change should be minimal, since PyTorch Lightning already deals itself to compile all internal functions (training_step, forward, ...).
The text was updated successfully, but these errors were encountered: