Update code format.

benchmarks/layer_benchmark/conv_benchmark.py

@@ -12,7 +12,6 @@
 ```
 """
 
-
 from absl import app
 from absl import flags
 

benchmarks/layer_benchmark/pooling_benchmark.py

@@ -12,7 +12,6 @@
 ```
 """
 
-
 from absl import app
 from absl import flags
 

benchmarks/torch_ctl_benchmark/conv_model_benchmark.py

@@ -4,6 +4,7 @@
 vanilla torch way, and we compare the performance between building model with
 Keras and torch.
 """
+
 import numpy as np
 import torch
 import torch.nn as nn

benchmarks/torch_ctl_benchmark/dense_model_benchmark.py

@@ -4,6 +4,7 @@
 vanilla torch way, and we compare the performance between building model with
 Keras and torch.
 """
+
 import numpy as np
 import torch
 import torch.nn as nn

guides/distributed_training_with_jax.py

@@ -6,6 +6,7 @@
 Description: Guide to multi-GPU/TPU training for Keras models with JAX.
 Accelerator: GPU
 """
+
 """
 ## Introduction
 

guides/distributed_training_with_tensorflow.py

@@ -6,6 +6,7 @@
 Description: Guide to multi-GPU training for Keras models with TensorFlow.
 Accelerator: GPU
 """
+
 """
 ## Introduction
 

guides/distributed_training_with_torch.py

@@ -6,6 +6,7 @@
 Description: Guide to multi-GPU training for Keras models with PyTorch.
 Accelerator: GPU
 """
+
 """
 ## Introduction
 

guides/functional_api.py

@@ -6,6 +6,7 @@
 Description: Complete guide to the functional API.
 Accelerator: GPU
 """
+
 """
 ## Setup
 """

guides/making_new_layers_and_models_via_subclassing.py

@@ -6,6 +6,7 @@
 Description: Complete guide to writing `Layer` and `Model` objects from scratch.
 Accelerator: None
 """
+
 """
 ## Introduction
 

guides/sequential_model.py

@@ -6,6 +6,7 @@
 Description: Complete guide to the Sequential model.
 Accelerator: GPU
 """
+
 """
 ## Setup
 

guides/transfer_learning.py

@@ -6,6 +6,7 @@
 Description: Complete guide to transfer learning & fine-tuning in Keras.
 Accelerator: GPU
 """
+
 """
 ## Setup
 """

guides/understanding_masking_and_padding.py

@@ -6,6 +6,7 @@
 Description: Complete guide to using mask-aware sequence layers in Keras.
 Accelerator: None
 """
+
 """
 ## Setup
 """

guides/writing_a_custom_training_loop_in_jax.py

@@ -6,6 +6,7 @@
 Description: Writing low-level training & evaluation loops in JAX.
 Accelerator: None
 """
+
 """
 ## Setup
 """

guides/writing_a_custom_training_loop_in_tensorflow.py

@@ -6,6 +6,7 @@
 Description: Writing low-level training & evaluation loops in TensorFlow.
 Accelerator: None
 """
+
 """
 ## Setup
 """

guides/writing_a_custom_training_loop_in_torch.py

@@ -6,6 +6,7 @@
 Description: Writing low-level training & evaluation loops in PyTorch.
 Accelerator: None
 """
+
 """
 ## Setup
 """

guides/writing_your_own_callbacks.py

@@ -6,6 +6,7 @@
 Description: Complete guide to writing new Keras callbacks.
 Accelerator: GPU
 """
+
 """
 ## Introduction
 

keras/backend/jax/distribution_lib.py

@@ -5,6 +5,7 @@
 This is just a prototype and we might want to unify it
 with other backends in the future.
 """
+
 import jax
 import numpy as np
 

keras/backend/numpy/math.py

@@ -16,9 +16,9 @@ def segment_sum(data, segment_ids, num_segments=None, sorted=False):
     valid_segment_ids = segment_ids[valid_indices]
 
     data_shape = list(valid_data.shape)
-    data_shape[
-        0
-    ] = num_segments  # Replace first dimension (which corresponds to segments)
+    data_shape[0] = (
+        num_segments  # Replace first dimension (which corresponds to segments)
+    )
 
     if sorted:
         result = np.zeros(data_shape, dtype=valid_data.dtype)
@@ -43,9 +43,9 @@ def segment_max(data, segment_ids, num_segments=None, sorted=False):
     valid_segment_ids = segment_ids[valid_indices]
 
     data_shape = list(valid_data.shape)
-    data_shape[
-        0
-    ] = num_segments  # Replace first dimension (which corresponds to segments)
+    data_shape[0] = (
+        num_segments  # Replace first dimension (which corresponds to segments)
+    )
 
     if sorted:
         result = np.zeros(data_shape, dtype=valid_data.dtype)

keras/backend/tensorflow/distribution_lib.py

@@ -5,6 +5,7 @@
 This is just a prototype and we might want to unify it
 with other backends in the future.
 """
+
 import tensorflow as tf
 from tensorflow.experimental import dtensor
 

keras/backend/tensorflow/numpy.py

@@ -576,9 +576,11 @@ def concatenate(xs, axis=0):
             return tf.sparse.concat(axis=axis, sp_inputs=xs)
         else:
             xs = [
-                convert_to_tensor(x, sparse=False)
-                if isinstance(x, tf.SparseTensor)
-                else x
+                (
+                    convert_to_tensor(x, sparse=False)
+                    if isinstance(x, tf.SparseTensor)
+                    else x
+                )
                 for x in xs
             ]
     xs = tf.nest.map_structure(convert_to_tensor, xs)

keras/backend/tensorflow/rnn.py

@@ -244,9 +244,11 @@ def _get_input_tensor(time):
             for i, inp in enumerate(flattened_inputs)
         )
         input_ta = tuple(
-            ta.unstack(input_)
-            if not go_backwards
-            else ta.unstack(tf.reverse(input_, [0]))
+            (
+                ta.unstack(input_)
+                if not go_backwards
+                else ta.unstack(tf.reverse(input_, [0]))
+            )
             for ta, input_ in zip(input_ta, flattened_inputs)
         )
 

keras/backend/torch/rnn.py

@@ -167,9 +167,11 @@ def _get_input_tensor(time):
         # flattened tensor.
 
         input_ta = tuple(
-            list(torch.unbind(input_))
-            if not go_backwards
-            else list(torch.unbind(torch.flip(input_, [0])))
+            (
+                list(torch.unbind(input_))
+                if not go_backwards
+                else list(torch.unbind(torch.flip(input_, [0])))
+            )
             for input_ in flattened_inputs
         )
 

keras/export/export_lib_test.py

@@ -1,4 +1,5 @@
 """Tests for inference-only model/layer exporting utilities."""
+
 import os
 
 import numpy as np

keras/layers/convolutional/base_separable_conv.py

@@ -1,6 +1,5 @@
 """Keras abstract base layer for separable convolution."""
 
-
 from keras import activations
 from keras import constraints
 from keras import initializers

keras/layers/convolutional/conv_test.py

@@ -206,9 +206,9 @@ def np_conv3d(
             (*new_kenel_size_tuple, ch_in, ch_out),
             dtype=kernel_weights.dtype,
         )
-        new_kernel_weights[
-            ::h_dilation, ::w_dilation, ::d_dilation
-        ] = kernel_weights
+        new_kernel_weights[::h_dilation, ::w_dilation, ::d_dilation] = (
+            kernel_weights
+        )
         kernel_weights = new_kernel_weights
         h_kernel, w_kernel, d_kernel = kernel_weights.shape[:3]
 

keras/layers/convolutional/conv_transpose_test.py

@@ -238,9 +238,9 @@ def np_conv3d_transpose(
             (*new_kenel_size_tuple, ch_out, ch_in),
             dtype=kernel_weights.dtype,
         )
-        new_kernel_weights[
-            ::h_dilation, ::w_dilation, ::d_dilation
-        ] = kernel_weights
+        new_kernel_weights[::h_dilation, ::w_dilation, ::d_dilation] = (
+            kernel_weights
+        )
         kernel_weights = new_kernel_weights
         h_kernel, w_kernel, d_kernel = kernel_weights.shape[:3]
 

keras/layers/core/einsum_dense_test.py

@@ -254,9 +254,9 @@ def test_einsum_dense_basics(
             },
             input_shape=input_shape,
             expected_output_shape=expected_output_shape,
-            expected_num_trainable_weights=2
-            if expected_bias_shape is not None
-            else 1,
+            expected_num_trainable_weights=(
+                2 if expected_bias_shape is not None else 1
+            ),
             expected_num_non_trainable_weights=0,
             expected_num_seed_generators=0,
             expected_num_losses=0,

keras/layers/core/lambda_layer.py

@@ -217,10 +217,10 @@ def from_config(cls, config, custom_objects=None, safe_mode=None):
                 )
                 config["output_shape"] = fn
             else:
-                config[
-                    "output_shape"
-                ] = serialization_lib.deserialize_keras_object(
-                    fn_config, custom_objects=custom_objects
+                config["output_shape"] = (
+                    serialization_lib.deserialize_keras_object(
+                        fn_config, custom_objects=custom_objects
+                    )
                 )
         if "arguments" in config:
             config["arguments"] = serialization_lib.deserialize_keras_object(

keras/layers/layer.py

@@ -15,6 +15,7 @@
 - RNG seed tracking
 - activity regularization
 """
+
 import collections
 import inspect
 import warnings

keras/layers/normalization/batch_normalization_test.py

@@ -132,9 +132,9 @@ def test_correctness(
         self.assertNotAllClose(unmasked_out, masked_out)
 
     @parameterized.product(
-        synchronized=(False, True)
-        if backend.backend == "tensorflow"
-        else (False,),
+        synchronized=(
+            (False, True) if backend.backend == "tensorflow" else (False,)
+        ),
     )
     def test_input_fully_masked(self, synchronized):
         norm = layers.BatchNormalization(

keras/layers/reshaping/cropping2d.py

@@ -103,12 +103,16 @@ def compute_output_shape(self, input_shape):
             return (
                 input_shape[0],
                 input_shape[1],
-                input_shape[2] - self.cropping[0][0] - self.cropping[0][1]
-                if input_shape[2] is not None
-                else None,
-                input_shape[3] - self.cropping[1][0] - self.cropping[1][1]
-                if input_shape[3] is not None
-                else None,
+                (
+                    input_shape[2] - self.cropping[0][0] - self.cropping[0][1]
+                    if input_shape[2] is not None
+                    else None
+                ),
+                (
+                    input_shape[3] - self.cropping[1][0] - self.cropping[1][1]
+                    if input_shape[3] is not None
+                    else None
+                ),
             )
         else:
             if (
@@ -125,12 +129,16 @@ def compute_output_shape(self, input_shape):
                 )
             return (
                 input_shape[0],
-                input_shape[1] - self.cropping[0][0] - self.cropping[0][1]
-                if input_shape[1] is not None
-                else None,
-                input_shape[2] - self.cropping[1][0] - self.cropping[1][1]
-                if input_shape[2] is not None
-                else None,
+                (
+                    input_shape[1] - self.cropping[0][0] - self.cropping[0][1]
+                    if input_shape[1] is not None
+                    else None
+                ),
+                (
+                    input_shape[2] - self.cropping[1][0] - self.cropping[1][1]
+                    if input_shape[2] is not None
+                    else None
+                ),
                 input_shape[3],
             )
 

keras/layers/reshaping/flatten_test.py

@@ -46,9 +46,11 @@ def test_flatten(self, sparse):
             init_kwargs={},
             input_data=inputs,
             input_sparse=True,
-            expected_output=expected_output_channels_last
-            if backend.config.image_data_format() == "channels_last"
-            else expected_output_channels_first,
+            expected_output=(
+                expected_output_channels_last
+                if backend.config.image_data_format() == "channels_last"
+                else expected_output_channels_first
+            ),
             expected_output_sparse=sparse,
             run_training_check=not sparse,
         )

keras/layers/rnn/conv_lstm1d_test.py

@@ -44,9 +44,9 @@ def test_basics(self):
                 "return_sequences": True,
             },
             input_shape=(3, 2, 8, 3) if channels_last else (3, 2, 3, 8),
-            expected_output_shape=(3, 2, 6, 5)
-            if channels_last
-            else (3, 2, 5, 6),
+            expected_output_shape=(
+                (3, 2, 6, 5) if channels_last else (3, 2, 5, 6)
+            ),
             expected_num_trainable_weights=3,
             expected_num_non_trainable_weights=0,
             supports_masking=True,