About solving Allen-Cahn equation

[alessio1233]

I trying to replicate the implementation of discrete time model of PINNs to solve the Allen-cahn equation. to do that, following the tutorial given, i set the initial condition and the boundary condition:

N0=200

mesh_s = pinnstorch.data.DiscreteMeshSampler(mesh = mesh,
idx_t=20,
num_sample=N0,
solution=['u'],
collection_points=['f'])

pe_b=pinnstorch.data.PeriodicBoundaryCondition(mesh=mesh,
idx_t=180,
derivative_order=1,
solution=['u'],
discrete=True)

val_s = pinnstorch.data.DiscreteMeshSampler(mesh = mesh,
idx_t=180,
solution = ['u'])

train_datasets = [mesh_s,pe_b]
val_dataset = val_s
datamodule = pinnstorch.data.PINNDataModule(train_datasets = train_datasets,
val_dataset = val_dataset,
pred_dataset = val_s)

Then, i set up the Runge-kutta method and the model:

net = pinnstorch.models.FCN(layers = [1, 200, 200, 200, 200, 101],
output_names=['u'],
lb=mesh.lb,
ub=mesh.ub,
discrete=True)

runge_kutta=pinnstorch.models.RungeKutta(root_dir='..',
t1=20,
t2=180,
time_domain=time_domain,
q=100)

model = pinnstorch.models.PINNModule(net = net,
pde_fn = pde_fn,
output_fn=None,
loss_fn='sse',
jit_compile=False,
runge_kutta=runge_kutta)

When i setu up trainer and i run the training, i get the following message:

C:\Users\userr\anaconda3\envs\lab\Lib\site-packages\lightning\pytorch\utilities\data.py:78: Trying to infer the batch_size from an ambiguous collection. The batch size we found is 512. To avoid any miscalculations, use self.log(..., batch_size=batch_size).

What's wrong whith my code? it seems to converge to real solution but, i cannot understand if the contranint are passed in the right way.

Thank you