Fix bugs in DataLoader and tutorial5 notebook

[FilippoOlivo]

No description provided.

[dario-coscia]

why are you not passing kwargs?

[FilippoOlivo]

kwargs are passed when TorchOptimizer is initialized outside the Solver class

from pina.optim import TorchOptimizer
from pina.solver import PINN
from torch.optim import SGD

optimizer = TorchOptimizer(SGD, lr=0.1, ...) 

in_ = LabelTensor(torch.rand((10,2)), ['u_0', 'u_1'])
out_ = LabelTensor(torch.rand((10,1)), ['u'])

class TestProblem(AbstractProblem):
    input_variables = ['u_0', 'u_1']
    output_variables = ['u']

    conditions = {
        'data': Condition(input_points=in_, output_points=out_),
    }

solver = SupervisedSolver(problem=problem, model=model, optimizer=optimizer)

[dario-coscia]

Can we pass the standard torch optimizer and inside solvers do the wrapper? To me, for the user, this is more intuitive

[dario-coscia]

@ndem0 what do you think?