Fix Keras build/call parameter naming in DeepSet layers

Author: SermetPekin

.gitignore

@@ -13,6 +13,8 @@ docsrc/source/contributing.md
 examples/checkpoints/
 build
 docs/
+.venv/
+.env
 
 
 # mypy

bayesflow/networks/deep_set/equivariant_layer.py

@@ -99,8 +99,8 @@ def __init__(
         self.layer_norm = layers.LayerNormalization() if layer_norm else None
 
     @sanitize_input_shape
-    def build(self, input_shape):
-        self.call(keras.ops.zeros(input_shape))
+    def build(self, input_set_shape):
+        self.call(keras.ops.zeros(input_set_shape))
 
     def call(self, input_set: Tensor, training: bool = False, **kwargs) -> Tensor:
         """Performs the forward pass of a learnable equivariant transform.

bayesflow/networks/deep_set/invariant_layer.py

@@ -115,5 +115,5 @@ def call(self, input_set: Tensor, training: bool = False, **kwargs) -> Tensor:
         return set_summary
 
     @sanitize_input_shape
-    def build(self, input_shape):
-        self.call(keras.ops.zeros(input_shape))
+    def build(self, input_set_shape):
+        self.call(keras.ops.zeros(input_set_shape))