The Helmholtz example

[ShaikhaTheGreen]

Hello everyone and thank you for this excellent library!

I'm trying this example: https://deepxde.readthedocs.io/en/latest/demos/pinn_forward/helmholtz.2d.neumann.hole.html. However the kernel crashes at this line:

losshistory, train_state = model.train(iterations=iterations)

There is not enough feedback to debug the problem. I only receive this message:

Canceled future for execute_request message before replies were done
The Kernel crashed while executing code in the the current cell or a previous cell. Please review the code in the cell(s) to identify a possible cause of the failure. Click here for more info. View Jupyter log for further details.

I have this line in a separate cell. This is how I noticed it was causing the problem. I'm also using the latest deepxde version, but I also tried version 1.2. All my attempts show the same results. I have also tried cutting down the training points to half of the number to inspect if it is a memory problem, but the error persisted.

Has anyone experienced this problem? Please assist.

[ShaikhaTheGreen]

Update:

I downgraded deepxde to 1.2 instead of 1.6. The same line also stops the run, but now it shows this error instead:

36 loss_weights = [1, weight_inner, weight_outer]
    37 model.compile(
    38     "adam", lr=learning_rate, metrics=["l2 relative error"], loss_weights=loss_weights
    39 )
---> 41 losshistory, train_state = model.train(epochs=iterations)
    42 dde.saveplot(losshistory, train_state, issave=True, isplot=True)

File c:\ProgramData\Anaconda3\envs\deepxde_1.2\lib\site-packages\deepxde\utils\internal.py:22, in timing.<locals>.wrapper(*args, **kwargs)
    19 @wraps(f)
    20 def wrapper(*args, **kwargs):
    21     ts = timeit.default_timer()
---> 22     result = f(*args, **kwargs)
    23     te = timeit.default_timer()
    24     print("%r took %f s\n" % (f.__name__, te - ts))

File c:\ProgramData\Anaconda3\envs\deepxde_1.2\lib\site-packages\deepxde\model.py:427, in Model.train(self, epochs, batch_size, display_every, disregard_previous_best, callbacks, model_restore_path, model_save_path)
   425 self.train_state.set_data_train(*self.data.train_next_batch(self.batch_size))
   426 self.train_state.set_data_test(*self.data.test())
--> 427 self._test()
   428 self.callbacks.on_train_begin()
   429 if optimizers.is_external_optimizer(self.opt_name):
...
   File "c:\ProgramData\Anaconda3\envs\deepxde_1.2\lib\site-packages\deepxde\icbc\boundary_conditions.py", line 54, in normal_derivative  *
       n = self.boundary_normal(X, beg, end)

   TypeError: tf__wrapper_nocache() missing 1 required positional argument: '_'

Any ideas?

[lululxvi]

@pescap

[pescap]

I'll have a try.

[pescap]

Hi @engsbk, which backend were you using?

[ShaikhaTheGreen]

I tried TF1 and TF2. On Sep 8, 2022, at 9:17 AM, pescapil ***@***.***> wrote:  Hi @engsbk<https://github.com/engsbk>, which backend were you using? — Reply to this email directly, view it on GitHub<#855 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AGBAUNRQ2CGZWOZ24PE3XZDV5HRPFANCNFSM57BDH3MA>. You are receiving this because you were mentioned.Message ID: ***@***.***>