add type hints for parameters and the return type

change Bilinear math to x_1 and x_2
change __call__ arguments to x and y instead of input and output
add explanation to the Initialization

python/mlx/nn/layers/linear.py

@@ -1,6 +1,7 @@
 # Copyright © 2023 Apple Inc.
 
 import math
+from typing import Any
 
 import mlx.core as mx
 from mlx.nn.layers.base import Module
@@ -14,11 +15,11 @@ class Identity(Module):
         kwargs: any keyword argument (unused)
     """
 
-    def __init__(self, *args, **kwargs):
+    def __init__(self, *args: Any, **kwargs: Any) -> None:
         super().__init__()
 
-    def __call__(self, input):
+    def __call__(self, x: mx.array) -> mx.array:
-        return input
+        return x
 
 
 class Linear(Module):
@@ -30,103 +31,94 @@ class Linear(Module):
 
         y = W^\top x + b
 
-    where :math:`W` has shape ``[output_dims, input_dims]``.
+    where:
+    :math:`W` has shape ``[output_dims, input_dims]``.
+    :math:`b` has shape ``[output_dims, ]``.
+
+    The values are initialized from :math:`\mathcal{U}(-{k}, {k})`, where
+    :math:`k = \frac{1}{\sqrt{input\_dims}}`.
 
     Args:
         input_dims (int): The dimensionality of the input features
         output_dims (int): The dimensionality of the output features
         bias (bool, optional): If set to ``False`` then the layer will
-          not use a bias. Default ``True``.
+          not use a bias. Default is ``True``.
     """
 
-    def __init__(self, input_dims: int, output_dims: int, bias: bool = True):
+    def __init__(self, input_dims: int, output_dims: int, bias: bool = True) -> None:
         super().__init__()
-        self.input_dims = input_dims
+        scale = math.sqrt(1.0 / input_dims)
-        self.output_dims = output_dims
-        self.weight = mx.zeros((output_dims, input_dims))
-        if bias:
-            self.bias = mx.zeros((output_dims,))
-
-        self.reset_parameters()
-
-    def reset_parameters(self):
-        scale = math.sqrt(1.0 / self.input_dims)
         self.weight = mx.random.uniform(
             low=-scale,
             high=scale,
-            shape=(self.output_dims, self.input_dims),
+            shape=(output_dims, input_dims),
         )
-        if "bias" in self:
+        if bias:
             self.bias = mx.random.uniform(
                 low=-scale,
                 high=scale,
-                shape=(self.output_dims,),
+                shape=(output_dims,),
             )
 
-    def _extra_repr(self):
+    def _extra_repr(self) -> str:
-        return f"input_dims={self.input_dims}, output_dims={self.output_dims}, bias={'bias' in self}"
+        return f"input_dims={self.weight.shape[1]}, output_dims={self.weight.shape[0]}, bias={'bias' in self}"
 
-    def __call__(self, input):
+    def __call__(self, x: mx.array) -> mx.array:
-        output = input @ self.weight.T
+        y = x @ self.weight.T
         if "bias" in self:
-            output = output + self.bias
+            y = y + self.bias
-        return output
+        return y
 
 
 class Bilinear(Module):
-    r"""Applies a bilinear transformation to the input.
+    r"""Applies a bilinear transformation to the inputs.
 
     Concretely:
 
     .. math::
 
-        y = input1^\top W input2 + b
+        y = x_1^\top W x_2 + b
 
-    where :math:`W` has shape ``[output_dims, input1_dims, input2_dims]``.
+    where
+    :math:`W` has shape ``[output_dims, input1_dims, input2_dims]``.
+    :math:`b` has shape ``[output_dims, ]``.
+
+    The values are initialized from :math:`\mathcal{U}(-{k}, {k})`, where
+    :math:`k = \frac{1}{\sqrt{input1\_dims}}`.
 
     Args:
         input1_dims (int): The dimensionality of the input1 features
         input2_dims (int): The dimensionality of the input2 features
         output_dims (int): The dimensionality of the output features
         bias (bool, optional): If set to ``False`` then the layer will
-          not use a bias. Default ``True``.
+          not use a bias. Default is ``True``.
     """
 
     def __init__(
         self, input1_dims: int, input2_dims: int, output_dims: int, bias: bool = True
-    ):
+    ) -> None:
         super().__init__()
-        self.input1_dims = input1_dims
+        scale = math.sqrt(1.0 / input1_dims)
-        self.input2_dims = input2_dims
-        self.output_dims = output_dims
-        self.weight = mx.zeros((output_dims, input1_dims, input2_dims))
-        if bias:
-            self.bias = mx.zeros((output_dims,))
-
-        self.reset_parameters()
-
-    def reset_parameters(self):
-        scale = math.sqrt(1.0 / self.input1_dims)
         self.weight = mx.random.uniform(
             low=-scale,
             high=scale,
-            shape=(self.output_dims, self.input1_dims, self.input2_dims),
+            shape=(output_dims, input1_dims, input2_dims),
         )
-        if "bias" in self:
+        if bias:
             self.bias = mx.random.uniform(
                 low=-scale,
                 high=scale,
-                shape=(self.output_dims,),
+                shape=(output_dims,),
             )
 
-    def _extra_repr(self):
+    def _extra_repr(self) -> str:
         return (
-            f"input1_dims={self.input1_dims}, input2_dims={self.input2_dims}, output_dims={self.output_dims}, "
+            f"input1_dims={self.weight.shape[1]}, input2_dims={self.weight.shape[2]}, "
-            f"bias={'bias' in self}"
+            f"output_dims={self.weight.shape[0]}, bias={'bias' in self}"
         )
 
-    def __call__(self, input1, input2):
+    def __call__(self, x1: mx.array, x2: mx.array) -> mx.array:
-        output = (input1 @ self.weight * input2.reshape(1, *input2.shape)).sum(-1).T
+        y = (x1 @ self.weight * x2.reshape(1, *x2.shape)).sum(-1).T
         if "bias" in self:
-            output = output + self.bias
+            y = y + self.bias
-        return output
+        return y