Make it runnable and fix runtime errors

src/base_trainer.py

@@ -98,12 +98,12 @@ def _train_val_iteration(
             torch.Tensor: The loss for the batch.
             Dict[str, torch.Tensor]: The loss components for the batch.
         """
-        # x, y = batch
-        # y_hat = self._model(x)
-        # losses = self._training_loss(x, y, y_hat)
-        # loss = sum([v for v in losses.values()])
-        # return loss, losses
-        raise NotImplementedError
+        # TODO: You'll most likely want to override this method.
+        x, y = batch
+        y_hat = self._model(x)
+        losses: Dict[str, torch.Tensor] = self._training_loss(y, y_hat)
+        loss = sum([v for v in losses.values()])
+        return loss, losses
 
     def _train_epoch(
         self, description: str, visualize: bool, epoch: int, last_val_loss: float
@@ -134,7 +134,7 @@ def _train_epoch(
                 break
             self._opt.zero_grad()
             loss, loss_components = self._train_val_iteration(
-                batch, epoch
+                batch, epoch, validation=False
             )  # User implementation goes here (train.py)
             loss.backward()
             self._opt.step()
@@ -193,7 +193,8 @@ def _val_epoch(self, description: str, visualize: bool, epoch: int) -> float:
                 # Blink the progress bar to indicate that the validation loop is running
                 blink_pbar(i, self._pbar, 4)
                 loss, loss_components = self._train_val_iteration(
-                    batch
+                    batch,
+                    epoch,
                 )  # User implementation goes here (train.py)
                 val_loss.update(loss.item())
                 for k, v in loss_components.items():

src/losses/mse.py

@@ -18,4 +18,8 @@ def __init__(self, reduction: str):
         self._reduction = reduction
 
     def __call__(self, y_pred: torch.Tensor, y_true: torch.Tensor) -> torch.Tensor:
-        return torch.nn.functional.mse_loss(y_pred, y_true, reduction=self._reduction)
+        return {
+            "mse": torch.nn.functional.mse_loss(
+                y_pred, y_true, reduction=self._reduction
+            )
+        }