diff --git a/lightgbmlss/distributions/distribution_utils.py b/lightgbmlss/distributions/distribution_utils.py
index 96a4173..a8e90ee 100644
--- a/lightgbmlss/distributions/distribution_utils.py
+++ b/lightgbmlss/distributions/distribution_utils.py
@@ -127,6 +127,7 @@ def metric_fn(self, predt: np.ndarray, data: lgb.Dataset) -> Tuple[str, float, b
         """
         # Target
         target = torch.tensor(data.get_label().reshape(-1, 1))
+        n_obs = target.shape[0]
 
         # Start values (needed to replace NaNs in predt)
         start_values = data.get_init_score().reshape(-1, self.n_dist_param)[0, :].tolist()
@@ -135,7 +136,7 @@ def metric_fn(self, predt: np.ndarray, data: lgb.Dataset) -> Tuple[str, float, b
         is_higher_better = False
         _, loss = self.get_params_loss(predt, target, start_values, requires_grad=False)
 
-        return self.loss_fn, loss, is_higher_better
+        return self.loss_fn, loss / n_obs, is_higher_better
 
     def loss_fn_start_values(self,
                              params: torch.Tensor,
diff --git a/lightgbmlss/distributions/flow_utils.py b/lightgbmlss/distributions/flow_utils.py
index b63d0a9..7b54272 100644
--- a/lightgbmlss/distributions/flow_utils.py
+++ b/lightgbmlss/distributions/flow_utils.py
@@ -143,6 +143,7 @@ def metric_fn(self, predt: np.ndarray, data: lgb.Dataset) -> Tuple[str, float, b
         """
         # Target
         target = torch.tensor(data.get_label().reshape(-1, 1))
+        n_obs = target.shape[0]
 
         # Start values (needed to replace NaNs in predt)
         start_values = data.get_init_score().reshape(-1, self.n_dist_param)[0, :].tolist()
@@ -151,7 +152,7 @@ def metric_fn(self, predt: np.ndarray, data: lgb.Dataset) -> Tuple[str, float, b
         is_higher_better = False
         _, loss = self.get_params_loss(predt, target, start_values)
 
-        return self.loss_fn, loss.detach(), is_higher_better
+        return self.loss_fn, loss.detach() / n_obs, is_higher_better
 
     def calculate_start_values(self,
                                target: np.ndarray,
diff --git a/lightgbmlss/distributions/mixture_distribution_utils.py b/lightgbmlss/distributions/mixture_distribution_utils.py
index 9d49360..ee26498 100644
--- a/lightgbmlss/distributions/mixture_distribution_utils.py
+++ b/lightgbmlss/distributions/mixture_distribution_utils.py
@@ -170,6 +170,7 @@ def metric_fn(self, predt: np.ndarray, data: lgb.Dataset) -> Tuple[str, float, b
         """
         # Target
         target = torch.tensor(data.get_label().reshape(-1, 1), dtype=torch.float32)
+        n_obs = target.shape[0]
 
         # Start values (needed to replace NaNs in predt)
         start_values = data.get_init_score().reshape(-1, self.n_dist_param)[0, :].tolist()
@@ -178,7 +179,7 @@ def metric_fn(self, predt: np.ndarray, data: lgb.Dataset) -> Tuple[str, float, b
         is_higher_better = False
         _, loss = self.get_params_loss(predt, target.flatten(), start_values, requires_grad=False)
 
-        return self.loss_fn, loss, is_higher_better
+        return self.loss_fn, loss / n_obs, is_higher_better
 
     def create_mixture_distribution(self,
                                     params: List[torch.Tensor],