diff --git a/pts/model/deepvar/deepvar_network.py b/pts/model/deepvar/deepvar_network.py
index 37cc7f3..c3a187d 100644
--- a/pts/model/deepvar/deepvar_network.py
+++ b/pts/model/deepvar/deepvar_network.py
@@ -234,7 +234,7 @@ class DeepVARTrainingNetwork(nn.Module):
         """
 
         past_observed_values = torch.min(
-            past_observed_values, past_is_pad.unsqueeze(-1)
+            past_observed_values, 1 - past_is_pad.unsqueeze(-1)
         )
 
         if future_time_feat is None or future_target_cdf is None:
diff --git a/pts/modules/distribution_output.py b/pts/modules/distribution_output.py
index 515a1a6..b4d2dee 100644
--- a/pts/modules/distribution_output.py
+++ b/pts/modules/distribution_output.py
@@ -203,7 +203,6 @@ class IndependentNormalOutput(DistributionOutput):
 class MultivariateNormalOutput(DistributionOutput):
     def __init__(self, dim: int) -> None:
         self.args_dim = {"loc": dim, "scale_tril": dim * dim}
-        self.distr_cls = MultivariateNormal
         self.dim = dim
 
     def domain_map(self, loc, scale):