diff --git a/src/continuiti/networks/attention.py b/src/continuiti/networks/attention.py
index 85421129..82e24dee 100644
--- a/src/continuiti/networks/attention.py
+++ b/src/continuiti/networks/attention.py
@@ -16,13 +16,10 @@ class Attention(nn.Module):
     components is designated using "soft" weights. These weights are assigned according to specific algorithms (e.g.
     scaled-dot-product attention).
 
-    Args:
-        dropout_p: dropout probability.
     """
 
-    def __init__(self, dropout_p: float = 0.0):
+    def __init__(self):
         super().__init__()
-        self.dropout_p = dropout_p
 
     @abstractmethod
     def forward(
diff --git a/src/continuiti/networks/scaled_dot_product_attention.py b/src/continuiti/networks/scaled_dot_product_attention.py
index a1d77cde..64f60bd3 100644
--- a/src/continuiti/networks/scaled_dot_product_attention.py
+++ b/src/continuiti/networks/scaled_dot_product_attention.py
@@ -19,7 +19,8 @@ class ScaledDotProductAttention(Attention):
     """
 
     def __init__(self, dropout_p: float = 0.0):
-        super().__init__(dropout_p)
+        super().__init__()
+        self.dropout_p = dropout_p
 
     def forward(
         self,