diff --git a/pts/model/deepar/deepar_estimator.py b/pts/model/deepar/deepar_estimator.py
index 5266aad..d2975b4 100644
--- a/pts/model/deepar/deepar_estimator.py
+++ b/pts/model/deepar/deepar_estimator.py
@@ -32,6 +32,7 @@ class DeepAREstimator(PTSEstimator):
             self,
             freq: str,
             prediction_length: int,
+            input_size: int,
             trainer: Trainer = Trainer(),
             context_length: Optional[int] = None,
             num_layers: int = 2,
@@ -58,6 +59,7 @@ class DeepAREstimator(PTSEstimator):
         self.prediction_length = prediction_length
         self.distr_output = distr_output
         self.distr_output.dtype = dtype
+        self.input_size = input_size
         self.num_layers = num_layers
         self.num_cells = num_cells
         self.cell_type = cell_type
@@ -65,9 +67,7 @@ class DeepAREstimator(PTSEstimator):
         self.use_feat_dynamic_real = use_feat_dynamic_real
         self.use_feat_static_cat = use_feat_static_cat
         self.use_feat_static_real = use_feat_static_real
-        self.cardinality = cardinality if cardinality and use_feat_static_cat else [
-            1
-        ]
+        self.cardinality = cardinality if cardinality and use_feat_static_cat else [1]
         self.embedding_dimension = (
             embedding_dimension if embedding_dimension is not None else
             [min(50, (cat + 1) // 2) for cat in self.cardinality])
@@ -153,6 +153,7 @@ class DeepAREstimator(PTSEstimator):
 
     def create_training_network(self, device: torch.device) -> DeepARTrainingNetwork:
         return DeepARTrainingNetwork(
+            input_size=self.input_size,
             num_layers=self.num_layers,
             num_cells=self.num_cells,
             cell_type=self.cell_type,
diff --git a/pts/model/deepar/deepar_network.py b/pts/model/deepar/deepar_network.py
index c5955fd..dabf715 100644
--- a/pts/model/deepar/deepar_network.py
+++ b/pts/model/deepar/deepar_network.py
@@ -17,6 +17,7 @@ def prod(xs):
 class DeepARNetwork(nn.Module):
     def __init__(
             self,
+            input_size: int,
             num_layers: int,
             num_cells: int,
             cell_type: str,
@@ -32,6 +33,7 @@ class DeepARNetwork(nn.Module):
             dtype: np.dtype = np.float32,
     ) -> None:
         super().__init__()
+        self.input_size = input_size
         self.num_layers = num_layers
         self.num_cells = num_cells
         self.cell_type = cell_type
@@ -49,7 +51,7 @@ class DeepARNetwork(nn.Module):
 
         self.distr_output = distr_output
         rnn = {"LSTM": nn.LSTM, "GRU": nn.GRU}[self.cell_type]
-        self.rnn = rnn(input_size=48,
+        self.rnn = rnn(input_size=input_size,
                        hidden_size=num_cells,
                        num_layers=num_layers,
                        dropout=dropout_rate,
@@ -72,8 +74,7 @@ class DeepARNetwork(nn.Module):
             sequence: torch.Tensor,
             sequence_length: int,
             indices: List[int],
-            subsequences_length: int = 1,
-    ) -> torch.Tensor:
+            subsequences_length: int = 1) -> torch.Tensor:
         """
         Returns lagged subsequences of a given sequence.
         Parameters
diff --git a/test/model/deepar/test_auxillary_outputs.py b/test/model/deepar/test_auxillary_outputs.py
index 2cdfeea..4d55a5b 100644
--- a/test/model/deepar/test_auxillary_outputs.py
+++ b/test/model/deepar/test_auxillary_outputs.py
@@ -35,6 +35,7 @@ def test_distribution():
     estimator = DeepAREstimator(
         freq=freq,
         prediction_length=prediction_length,
+        input_size=48,
         trainer=Trainer(epochs=1, num_batches_per_epoch=1),
         distr_output=StudentTOutput(),
     )