Hello! It seems the 'targets' variable in the model.py module, which, when used with the pde.py module corresponds to the 'solutions' argument of the 'PDE' class (I think), is not actually being taken into account in the 'losses()' method of pde.py.
Consequently, none of the methods starting from 'outputs_losses()' all the way up to 'train()' and '_test()' in model.py take into account any reference data. Is that right? I may be mistaken as I am not very experienced with ML just yet. If not, could you please explain what the 'targets' variable is? Issue #358 explored this question, but I don't believe it was ever answered.
Hello! It seems the 'targets' variable in the model.py module, which, when used with the pde.py module corresponds to the 'solutions' argument of the 'PDE' class (I think), is not actually being taken into account in the 'losses()' method of pde.py.
Consequently, none of the methods starting from 'outputs_losses()' all the way up to 'train()' and '_test()' in model.py take into account any reference data. Is that right? I may be mistaken as I am not very experienced with ML just yet. If not, could you please explain what the 'targets' variable is? Issue #358 explored this question, but I don't believe it was ever answered.
Thanks!