Prakhar Sharma

Results 85 comments of Prakhar Sharma

What is adaptive weight template of deepxde?

One of my friend is doing the same thing with [GPR](https://scikit-learn.org/stable/modules/gaussian_process.html#gaussian-process-regression-gpr). But it is very hard to implement this in PINNs without the reconstruction loss. I also trying to do...

Which Tensorflow version are you using? Use Tensorflow 2.x.

TLDR. If your loss is converging and you are getting a wrong solution to the PDE then it means that either the problem is ill-posed i.e. there are multiple solutions...

What do you mean when you say " And there seems to be no solution to this problem as far as I know"? The problem has no solution when you...

This is the original paper. You can find a Python code by searching for citations on Google scholar. [Original_paper.pdf](https://github.com/lululxvi/deepxde/files/9504439/Original_paper.pdf) I can see your code. It would be better if you...

Ok. Let us start from basics. When You solve a PDE using FEM, you do not supply the true solution. Because if you already know the solution why would you...

Will look into your problem. Why do you need two rectangles `rectanglar2 = dde.geometry.Rectangle(xmin=[2.49,3], xmax=[2.51,3.1])`? Yes in some PINN frameworks people use the true solution during the training. We refer...

You can apply the reconstruction error using `PointSetBC`. [Here](https://github.com/lululxvi/deepxde/blob/e95d983d154ef0543f4bf0bf467b3d0ee3b13fdb/deepxde/icbc/boundary_conditions.py#L153) is the source code.

Yes it is possible and that why I told you about `PointSetBC`. This is how you use it ```python reconstruction_constraint = dde.icbc.PointSetBC(true_x, true_y) ``` where `true_x` and `true_y` are simply...