Remove outlet

src/nse/controller/simulation.py

@@ -20,22 +20,18 @@ def __init__(self, experiment: Experiment, device: str, steps: int, layers: list
         knowledge = self.__experiment.knowledge.detach()
         learning = self.__experiment.learning.detach()
         inlet = self.__experiment.inlet.detach()
-        outlet = self.__experiment.outlet.detach()
 
         self.__u = torch.tensor([[i.u] for _, i in inlet + knowledge], dtype=torch.float64, device=self._device)
         self.__v = torch.tensor([[i.v] for _, i in inlet + knowledge], dtype=torch.float64, device=self._device)
 
         self.__null = torch.zeros(len(learning), 1, dtype=torch.float64, device=self._device)
 
-        self.__out = len(outlet)
-        self.__u_out = torch.zeros(len(outlet), 1, dtype=torch.float64, device=self._device)
-
         self.__knowledge = (
-            torch.tensor([[k.x] for k, _ in outlet + inlet + knowledge],
+            torch.tensor([[k.x] for k, _ in inlet + knowledge],
                          dtype=torch.float64,
                          requires_grad=True,
                          device=self._device),
-            torch.tensor([[k.y] for k, _ in outlet + inlet + knowledge],
+            torch.tensor([[k.y] for k, _ in inlet + knowledge],
                          dtype=torch.float64,
                          requires_grad=True,
                          device=self._device),
@@ -91,12 +87,8 @@ def __loss(self):
 
         u, v, *_ = self.__forward(self.__knowledge)
 
-        u_loss = self._mse(u[self.__out:], self.__u)
-        v_loss = self._mse(v[self.__out:], self.__v)
-
-        # # prohibits the model from hallucinating an incoming flow from right
-        # if self.__out > 0:
-        #     u_loss += self._mse(torch.clamp(u[:self.__out], max=0), self.__u_out)
+        u_loss = self._mse(u, self.__u)
+        v_loss = self._mse(v, self.__v)
 
         *_, f, g = self.__forward(self.__learning, True)