Correctly set output_size when hidden_cell=None

nengo · Apr 19, 2024 · 330856d · 330856d
1 parent 4d7fe22
commit 330856d
Showing 1 changed file with 18 additions and 0 deletions.
diff --git a/keras_lmu/layers.py b/keras_lmu/layers.py
@@ -118,6 +118,7 @@ def __init__(
         order,
         theta,
         hidden_cell,
+        input_d=None,
         trainable_theta=False,
         hidden_to_memory=False,
         memory_to_memory=False,
@@ -141,6 +142,7 @@ def __init__(
         self.order = order
         self._init_theta = theta
         self.hidden_cell = hidden_cell
+        self.input_d = input_d
         self.trainable_theta = trainable_theta
         self.hidden_to_memory = hidden_to_memory
         self.memory_to_memory = memory_to_memory
@@ -178,6 +180,15 @@ def __init__(
                 )
 
             self.hidden_output_size = self.memory_d * self.order
+
+            if self.input_to_hidden:
+                if self.input_d is None:
+                    raise ValueError(
+                        "input_d must be specified when setting input_to_hidden=True "
+                        "with hidden_cell=None"
+                    )
+                self.hidden_output_size += self.input_d
+
             self.hidden_state_size = []
         elif hasattr(self.hidden_cell, "state_size"):
             self.hidden_output_size = self.hidden_cell.output_size
@@ -272,6 +283,12 @@ def build(self, input_shape):
 
         super().build(input_shape)
 
+        if self.input_d is not None and input_shape[-1] != self.input_d:
+            raise ValueError(
+                f"Input dimensionality ({input_shape[-1]}) does not match expected "
+                f"dimensionality ({self.input_d})"
+            )
+
         enc_d = input_shape[-1]
         if self.hidden_to_memory:
             enc_d += self.hidden_output_size
@@ -470,6 +487,7 @@ def get_config(self):
                 "order": self.order,
                 "theta": self._init_theta,
                 "hidden_cell": keras.layers.serialize(self.hidden_cell),
+                "input_d": self.input_d,
                 "trainable_theta": self.trainable_theta,
                 "hidden_to_memory": self.hidden_to_memory,
                 "memory_to_memory": self.memory_to_memory,