tf-keras-nightly 2.18.0.dev2024091909__tar.gz → 2.19.0.dev2024092709__tar.gz

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: tf_keras-nightly
-Version: 2.18.0.dev2024091909
+Version: 2.19.0.dev2024092709
 Summary: Deep learning for humans.
 Home-page: https://keras.io/
 Download-URL: https://github.com/keras-team/tf-keras/tags
@@ -26,7 +26,7 @@ Classifier: Topic :: Software Development
 Classifier: Topic :: Software Development :: Libraries
 Classifier: Topic :: Software Development :: Libraries :: Python Modules
 Requires-Python: >=3.9
-Requires-Dist: tf-nightly~=2.18.0.dev
+Requires-Dist: tf-nightly~=2.19.0.dev
 
 TF-Keras is a deep learning API written in Python,
 running on top of the machine learning platform TensorFlow.

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/setup.py

@@ -38,7 +38,7 @@ with open(os.path.abspath(__file__)) as f:
         )
 
 # pin version to that of tensorflow or tf_nightly.
-version = "2.18.0.dev2024091909".lower()
+version = "2.19.0.dev2024092709".lower()
 major_version, minor_version, *_ = version.split(".")
 next_minor_version = int(minor_version) + 1
 if "nightly" in "tf_keras-nightly":
@@ -61,7 +61,7 @@ setuptools.setup(
     name="tf_keras-nightly",
     # Version strings with `-` characters are semver compatible,
     # but incompatible with pip. For pip, we will remove all `-`` characters.
-    version="2.18.0.dev2024091909",
+    version="2.19.0.dev2024092709",
     description="Deep learning for humans.",
     long_description=DESCRIPTION,
     url="https://keras.io/",

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/__init__.py

@@ -27,4 +27,4 @@ from tf_keras.src.engine.sequential import Sequential
 from tf_keras.src.engine.training import Model
 
 
-__version__ = "2.18.0.dev2024091909"
+__version__ = "2.19.0.dev2024092709"

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/__init__.py

@@ -35,7 +35,7 @@ from tf_keras.src.testing_infra import test_utils
 from tensorflow.python import tf2
 from tensorflow.python.util.tf_export import keras_export
 
-__version__ = "2.18.0"
+__version__ = "2.19.0"
 
 keras_export("keras.__version__").export_constant(__name__, "__version__")
 

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/activation/prelu.py

@@ -102,7 +102,7 @@ class PReLU(Layer):
                 if i not in self.shared_axes:
                     axes[i] = input_shape[i]
         self.input_spec = InputSpec(ndim=len(input_shape), axes=axes)
-        self.built = True
+        super().build(input_shape)
 
     def call(self, inputs):
         pos = backend.relu(inputs)

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/attention/base_dense_attention.py

@@ -86,7 +86,8 @@ class BaseDenseAttention(base_layer.BaseRandomLayer):
         # be purely stateless, with no reference to any variable.
         if self.dropout > 0:
             super().build(input_shape)
-        self.built = True
+        else:
+            base_layer.Layer.build(self, input_shape)
 
     def _calculate_scores(self, query, key):
         """Calculates attention scores.

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/convolutional/base_conv.py

@@ -248,7 +248,7 @@ class Conv(Layer):
         self.input_spec = InputSpec(
             min_ndim=self.rank + 2, axes={channel_axis: input_channel}
         )
-        self.built = True
+        super().build(input_shape)
 
     def convolution_op(self, inputs, kernel):
         if self.padding == "causal":

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/convolutional/base_depthwise_conv.py

@@ -20,6 +20,7 @@ import tensorflow.compat.v2 as tf
 from tf_keras.src import constraints
 from tf_keras.src import initializers
 from tf_keras.src import regularizers
+from tf_keras.src.engine.base_layer import Layer
 from tf_keras.src.engine.input_spec import InputSpec
 from tf_keras.src.layers.convolutional.base_conv import Conv
 
@@ -202,7 +203,8 @@ class DepthwiseConv(Conv):
         self.input_spec = InputSpec(
             min_ndim=self.rank + 2, axes={channel_axis: input_dim}
         )
-        self.built = True
+        # Call Layer.build() to skip Conv.build() which we override here.
+        Layer.build(self, input_shape)
 
     def call(self, inputs):
         raise NotImplementedError

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/convolutional/base_separable_conv.py

@@ -21,6 +21,7 @@ from tf_keras.src import activations
 from tf_keras.src import constraints
 from tf_keras.src import initializers
 from tf_keras.src import regularizers
+from tf_keras.src.engine.base_layer import Layer
 from tf_keras.src.engine.input_spec import InputSpec
 from tf_keras.src.layers.convolutional.base_conv import Conv
 
@@ -203,7 +204,8 @@ class SeparableConv(Conv):
             )
         else:
             self.bias = None
-        self.built = True
+        # Call Layer.build() to skip Conv.build() which we override here.
+        Layer.build(self, input_shape)
 
     def call(self, inputs):
         raise NotImplementedError

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/convolutional/conv1d_transpose.py

@@ -22,6 +22,7 @@ from tf_keras.src import constraints
 from tf_keras.src import initializers
 from tf_keras.src import regularizers
 from tf_keras.src.dtensor import utils
+from tf_keras.src.engine.base_layer import Layer
 from tf_keras.src.engine.input_spec import InputSpec
 from tf_keras.src.layers.convolutional.conv1d import Conv1D
 from tf_keras.src.utils import conv_utils
@@ -214,7 +215,8 @@ class Conv1DTranspose(Conv1D):
             )
         else:
             self.bias = None
-        self.built = True
+        # Call Layer.build() to skip Conv.build() which we override here.
+        Layer.build(self, input_shape)
 
     def call(self, inputs):
         inputs_shape = tf.shape(inputs)

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/convolutional/conv2d_transpose.py

@@ -23,6 +23,7 @@ from tf_keras.src import constraints
 from tf_keras.src import initializers
 from tf_keras.src import regularizers
 from tf_keras.src.dtensor import utils
+from tf_keras.src.engine.base_layer import Layer
 from tf_keras.src.engine.input_spec import InputSpec
 from tf_keras.src.layers.convolutional.conv2d import Conv2D
 from tf_keras.src.utils import conv_utils
@@ -240,7 +241,8 @@ class Conv2DTranspose(Conv2D):
             )
         else:
             self.bias = None
-        self.built = True
+        # Call Layer.build() to skip Conv.build() which we override here.
+        Layer.build(self, input_shape)
 
     def call(self, inputs):
         inputs_shape = tf.shape(inputs)

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/convolutional/conv3d_transpose.py

@@ -22,6 +22,7 @@ from tf_keras.src import constraints
 from tf_keras.src import initializers
 from tf_keras.src import regularizers
 from tf_keras.src.dtensor import utils
+from tf_keras.src.engine.base_layer import Layer
 from tf_keras.src.engine.input_spec import InputSpec
 from tf_keras.src.layers.convolutional.conv3d import Conv3D
 from tf_keras.src.utils import conv_utils
@@ -247,7 +248,8 @@ class Conv3DTranspose(Conv3D):
             )
         else:
             self.bias = None
-        self.built = True
+        # Call Layer.build() to skip Conv.build() which we override here.
+        Layer.build(self, input_shape)
 
     def call(self, inputs):
         inputs_shape = tf.shape(inputs)

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/core/dense.py

@@ -174,7 +174,7 @@ class Dense(Layer):
             )
         else:
             self.bias = None
-        self.built = True
+        super().build(input_shape)
 
     def call(self, inputs):
         if inputs.dtype.base_dtype != self._compute_dtype_object.base_dtype:

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/core/embedding.py

@@ -185,7 +185,7 @@ class Embedding(Layer):
             constraint=self.embeddings_constraint,
             experimental_autocast=False,
         )
-        self.built = True
+        super().build(input_shape)
 
     def compute_mask(self, inputs, mask=None):
         if not self.mask_zero:

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/locally_connected/locally_connected1d.py

@@ -284,7 +284,7 @@ class LocallyConnected1D(Layer):
             self.input_spec = InputSpec(ndim=3, axes={1: input_dim})
         else:
             self.input_spec = InputSpec(ndim=3, axes={-1: input_dim})
-        self.built = True
+        super().build(input_shape)
 
     @tf_utils.shape_type_conversion
     def compute_output_shape(self, input_shape):

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/locally_connected/locally_connected2d.py

@@ -308,7 +308,7 @@ class LocallyConnected2D(Layer):
             self.input_spec = InputSpec(ndim=4, axes={1: input_filter})
         else:
             self.input_spec = InputSpec(ndim=4, axes={-1: input_filter})
-        self.built = True
+        super().build(input_shape)
 
     @tf_utils.shape_type_conversion
     def compute_output_shape(self, input_shape):

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/normalization/batch_normalization.py

@@ -542,7 +542,7 @@ class BatchNormalizationBase(Layer):
         finally:
             if partitioner:
                 self._scope.set_partitioner(partitioner)
-        self.built = True
+        super().build(input_shape)
 
     def call(self, inputs, training=None, mask=None):
         inputs = tf.cast(inputs, self.compute_dtype)

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/normalization/layer_normalization.py

@@ -249,7 +249,7 @@ class LayerNormalization(Layer):
             self.beta = None
 
         self._fused = self._fused_can_be_used(rank)
-        self.built = True
+        super().build(input_shape)
 
     def call(self, inputs):
         # TODO(b/229545225): Remove the RaggedTensor check.

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/rnn/abstract_rnn_cell.py

@@ -56,7 +56,7 @@ class AbstractRNNCell(base_layer.Layer):
               shape=(self.units, self.units),
               initializer='uniform',
               name='recurrent_kernel')
-          self.built = True
+          super().build(input_shape)
 
         def call(self, inputs, states):
           prev_output = states[0]

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/rnn/base_conv_lstm.py

@@ -218,7 +218,6 @@ class ConvLSTMCell(DropoutRNNCellMixin, base_layer.BaseRandomLayer):
             )
         else:
             self.bias = None
-        self.built = True
 
     def call(self, inputs, states, training=None):
         h_tm1 = states[0]  # previous memory state

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/rnn/base_conv_rnn.py

@@ -20,6 +20,7 @@ import tensorflow.compat.v2 as tf
 
 from tf_keras.src import backend
 from tf_keras.src.engine import base_layer
+from tf_keras.src.engine.base_layer import Layer
 from tf_keras.src.engine.input_spec import InputSpec
 from tf_keras.src.layers.rnn.base_rnn import RNN
 from tf_keras.src.utils import conv_utils
@@ -207,6 +208,8 @@ class ConvRNN(RNN):
 
     @tf_utils.shape_type_conversion
     def build(self, input_shape):
+        # Call Layer.build() to skip RNN.build() which we override here.
+        Layer.build(self, input_shape)
         # Note input_shape will be list of shapes of initial states and
         # constants if these are passed in __call__.
         if self._num_constants is not None:
@@ -263,7 +266,6 @@ class ConvRNN(RNN):
                 ]
         if self.stateful:
             self.reset_states()
-        self.built = True
 
     def get_initial_state(self, inputs):
         # (samples, timesteps, img_dims..., filters)

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/rnn/base_rnn.py

@@ -207,7 +207,7 @@ class RNN(base_layer.Layer):
                 shape=(self.units, self.units),
                 initializer='uniform',
                 name='recurrent_kernel')
-            self.built = True
+            super().build(input_shape)
 
         def call(self, inputs, states):
             prev_output = states[0]

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/rnn/base_wrapper.py

@@ -56,7 +56,7 @@ class Wrapper(Layer):
         if not self.layer.built:
             self.layer.build(input_shape)
             self.layer.built = True
-        self.built = True
+        super().build(input_shape)
 
     @property
     def activity_regularizer(self):

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/rnn/bidirectional.py

@@ -470,7 +470,8 @@ class Bidirectional(Wrapper):
             self.forward_layer.build(input_shape)
         with backend.name_scope(self.backward_layer.name):
             self.backward_layer.build(input_shape)
-        self.built = True
+        # Call Layer.build() to skip Wrapper.build() which we override here.
+        Layer.build(self, input_shape)
 
     def compute_mask(self, inputs, mask):
         if isinstance(mask, list):

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/rnn/cell_wrappers.py

@@ -102,10 +102,10 @@ class _RNNCellWrapper(AbstractRNNCell):
             inputs, state, cell_call_fn=self.cell.call, **kwargs
         )
 
-    def build(self, inputs_shape):
+    def build(self, input_shape):
         """Builds the wrapped cell."""
-        self.cell.build(inputs_shape)
-        self.built = True
+        self.cell.build(input_shape)
+        super().build(input_shape)
 
     @property
     def wrapped_cell(self):

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/rnn/cudnn_gru.py

@@ -144,8 +144,6 @@ class CuDNNGRU(_CuDNNRNN):
             constraint=self.bias_constraint,
         )
 
-        self.built = True
-
     def _process_batch(self, inputs, initial_state):
         if not self.time_major:
             inputs = tf.transpose(inputs, perm=(1, 0, 2))

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/rnn/cudnn_lstm.py

@@ -170,8 +170,6 @@ class CuDNNLSTM(_CuDNNRNN):
             constraint=self.bias_constraint,
         )
 
-        self.built = True
-
     def _process_batch(self, inputs, initial_state):
         if not self.time_major:
             inputs = tf.transpose(inputs, perm=(1, 0, 2))

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/rnn/gru.py

@@ -222,7 +222,6 @@ class GRUCell(DropoutRNNCellMixin, base_layer.BaseRandomLayer):
             )
         else:
             self.bias = None
-        self.built = True
 
     def call(self, inputs, states, training=None):
         h_tm1 = (

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/rnn/legacy_cell_wrappers.py

@@ -368,9 +368,9 @@ class DropoutWrapper(_RNNCellWrapperV1):
     def wrapped_cell(self):
         return self.cell
 
-    def build(self, inputs_shape):
-        self.cell.build(inputs_shape)
-        self.built = True
+    def build(self, input_shape):
+        self.cell.build(input_shape)
+        super().build(input_shape)
 
     def _variational_recurrent_dropout_value(
         self, unused_index, value, noise, keep_prob

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/rnn/legacy_cells.py

@@ -246,11 +246,6 @@ class RNNCell(base_layer.Layer):
         """Integer or TensorShape: size of outputs produced by this cell."""
         raise NotImplementedError("Abstract method")
 
-    def build(self, _):
-        # This tells the parent Layer object that it's OK to call
-        # self.add_weight() inside the call() method.
-        pass
-
     def get_initial_state(self, inputs=None, batch_size=None, dtype=None):
         if inputs is not None:
             # Validate the given batch_size and dtype against inputs if
@@ -445,15 +440,15 @@ class BasicRNNCell(LayerRNNCell):
         return self._num_units
 
     @tf_utils.shape_type_conversion
-    def build(self, inputs_shape):
-        if inputs_shape[-1] is None:
+    def build(self, input_shape):
+        if input_shape[-1] is None:
             raise ValueError(
                 "Expected inputs.shape[-1] to be known, "
-                f"received shape: {inputs_shape}"
+                f"received shape: {input_shape}"
             )
         _check_supported_dtypes(self.dtype)
 
-        input_depth = inputs_shape[-1]
+        input_depth = input_shape[-1]
         self._kernel = self.add_weight(
             _WEIGHTS_VARIABLE_NAME,
             shape=[input_depth + self._num_units, self._num_units],
@@ -464,7 +459,7 @@ class BasicRNNCell(LayerRNNCell):
             initializer=tf.compat.v1.zeros_initializer(dtype=self.dtype),
         )
 
-        self.built = True
+        super().build(input_shape)
 
     def call(self, inputs, state):
         """Most basic RNN: output = new_state = act(W * input + U * state +
@@ -563,14 +558,14 @@ class GRUCell(LayerRNNCell):
         return self._num_units
 
     @tf_utils.shape_type_conversion
-    def build(self, inputs_shape):
-        if inputs_shape[-1] is None:
+    def build(self, input_shape):
+        if input_shape[-1] is None:
             raise ValueError(
                 "Expected inputs.shape[-1] to be known, "
-                f"received shape: {inputs_shape}"
+                f"received shape: {input_shape}"
             )
         _check_supported_dtypes(self.dtype)
-        input_depth = inputs_shape[-1]
+        input_depth = input_shape[-1]
         self._gate_kernel = self.add_weight(
             f"gates/{_WEIGHTS_VARIABLE_NAME}",
             shape=[input_depth + self._num_units, 2 * self._num_units],
@@ -600,7 +595,7 @@ class GRUCell(LayerRNNCell):
             ),
         )
 
-        self.built = True
+        super().build(input_shape)
 
     def call(self, inputs, state):
         """Gated recurrent unit (GRU) with nunits cells."""
@@ -774,14 +769,14 @@ class BasicLSTMCell(LayerRNNCell):
         return self._num_units
 
     @tf_utils.shape_type_conversion
-    def build(self, inputs_shape):
-        if inputs_shape[-1] is None:
+    def build(self, input_shape):
+        if input_shape[-1] is None:
             raise ValueError(
                 "Expected inputs.shape[-1] to be known, "
-                f"received shape: {inputs_shape}"
+                f"received shape: {input_shape}"
             )
         _check_supported_dtypes(self.dtype)
-        input_depth = inputs_shape[-1]
+        input_depth = input_shape[-1]
         h_depth = self._num_units
         self._kernel = self.add_weight(
             _WEIGHTS_VARIABLE_NAME,
@@ -793,7 +788,7 @@ class BasicLSTMCell(LayerRNNCell):
             initializer=tf.compat.v1.zeros_initializer(dtype=self.dtype),
         )
 
-        self.built = True
+        super().build(input_shape)
 
     def call(self, inputs, state):
         """Long short-term memory cell (LSTM).
@@ -1017,14 +1012,14 @@ class LSTMCell(LayerRNNCell):
         return self._output_size
 
     @tf_utils.shape_type_conversion
-    def build(self, inputs_shape):
-        if inputs_shape[-1] is None:
+    def build(self, input_shape):
+        if input_shape[-1] is None:
             raise ValueError(
                 "Expected inputs.shape[-1] to be known, "
-                f"received shape: {inputs_shape}"
+                f"received shape: {input_shape}"
             )
         _check_supported_dtypes(self.dtype)
-        input_depth = inputs_shape[-1]
+        input_depth = input_shape[-1]
         h_depth = self._num_units if self._num_proj is None else self._num_proj
         maybe_partitioner = (
             tf.compat.v1.fixed_size_partitioner(self._num_unit_shards)
@@ -1076,7 +1071,7 @@ class LSTMCell(LayerRNNCell):
                 partitioner=maybe_proj_partitioner,
             )
 
-        self.built = True
+        super().build(input_shape)
 
     def call(self, inputs, state):
         """Run one step of LSTM.

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/rnn/lstm.py

@@ -236,7 +236,6 @@ class LSTMCell(DropoutRNNCellMixin, base_layer.BaseRandomLayer):
             )
         else:
             self.bias = None
-        self.built = True
 
     def _compute_carry_and_output(self, x, h_tm1, c_tm1):
         """Computes carry and output using split kernels."""

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/rnn/simple_rnn.py

@@ -189,7 +189,6 @@ class SimpleRNNCell(DropoutRNNCellMixin, base_layer.BaseRandomLayer):
             )
         else:
             self.bias = None
-        self.built = True
 
     def call(self, inputs, states, training=None):
         prev_output = states[0] if tf.nest.is_nested(states) else states

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/rnn/stacked_rnn_cells.py

@@ -166,6 +166,7 @@ class StackedRNNCells(base_layer.Layer):
 
     @tf_utils.shape_type_conversion
     def build(self, input_shape):
+        super().build(input_shape)
         if isinstance(input_shape, list):
             input_shape = input_shape[0]
 
@@ -195,7 +196,6 @@ class StackedRNNCells(base_layer.Layer):
                 input_shape = tuple(
                     [batch_size] + tf.TensorShape(output_dim).as_list()
                 )
-        self.built = True
 
     def get_config(self):
         cells = []

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/layers/rnn/time_distributed.py

@@ -135,7 +135,6 @@ class TimeDistributed(Wrapper):
         )
         child_input_shape = tf_utils.convert_shapes(child_input_shape)
         super().build(tuple(child_input_shape))
-        self.built = True
 
     def compute_output_shape(self, input_shape):
         input_shape = tf_utils.convert_shapes(input_shape, to_tuples=False)

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/mixed_precision/test_util.py

@@ -171,14 +171,14 @@ class MultiplyLayer(AssertTypeLayer):
             activity_regularizer=self._activity_regularizer, **kwargs
         )
 
-    def build(self, _):
+    def build(self, input_shape):
         self.v = self.add_weight(
             self._var_name,
             (),
             initializer="ones",
             regularizer=self._regularizer,
         )
-        self.built = True
+        super().build(input_shape)
 
     def call(self, inputs):
         self.assert_input_types(inputs)
@@ -205,7 +205,7 @@ class MultiplyLayer(AssertTypeLayer):
 class MultiplyLayerWithoutAutoCast(MultiplyLayer):
     """Same as MultiplyLayer, but does not use AutoCastVariables."""
 
-    def build(self, _):
+    def build(self, input_shape):
         dtype = self.dtype
         if dtype in ("float16", "bfloat16"):
             dtype = "float32"
@@ -217,7 +217,8 @@ class MultiplyLayerWithoutAutoCast(MultiplyLayer):
             autocast=False,
             regularizer=self._regularizer,
         )
-        self.built = True
+        # Call Layer.build() to skip MultiplyLayer.build() which we override.
+        base_layer.Layer.build(self, input_shape)
 
     def call(self, inputs):
         self.assert_input_types(inputs)

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras/src/premade_models/linear.py

@@ -156,7 +156,8 @@ class LinearModel(training.Model):
             )
         else:
             self.bias = None
-        self.built = True
+        # Call Layer.build() to skip Model.build() which we override here.
+        base_layer.Layer.build(self, input_shape)
 
     def call(self, inputs):
         result = None

{tf_keras_nightly-2.18.0.dev2024091909 → tf_keras_nightly-2.19.0.dev2024092709}/tf_keras_nightly.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: tf_keras-nightly
-Version: 2.18.0.dev2024091909
+Version: 2.19.0.dev2024092709
 Summary: Deep learning for humans.
 Home-page: https://keras.io/
 Download-URL: https://github.com/keras-team/tf-keras/tags
@@ -26,7 +26,7 @@ Classifier: Topic :: Software Development
 Classifier: Topic :: Software Development :: Libraries
 Classifier: Topic :: Software Development :: Libraries :: Python Modules
 Requires-Python: >=3.9
-Requires-Dist: tf-nightly~=2.18.0.dev
+Requires-Dist: tf-nightly~=2.19.0.dev
 
 TF-Keras is a deep learning API written in Python,
 running on top of the machine learning platform TensorFlow.

tf_keras_nightly-2.19.0.dev2024092709/tf_keras_nightly.egg-info/requires.txt

@@ -0,0 +1 @@
+tf-nightly~=2.19.0.dev

tf_keras_nightly-2.18.0.dev2024091909/tf_keras_nightly.egg-info/requires.txt

@@ -1 +0,0 @@
-tf-nightly~=2.18.0.dev