Standardize docstring (#2399)

benchmarks/vectorized_channel_shuffle.py

@@ -50,7 +50,7 @@ class OldChannelShuffle(BaseImageAugmentationLayer):
         training: A boolean argument that determines whether the call should be
             run in inference mode or training mode, defaults to True.
 
-    Usage:
+    Example:
     ```python
     (images, labels), _ = keras.datasets.cifar10.load_data()
     channel_shuffle = keras_cv.layers.ChannelShuffle()

benchmarks/vectorized_grayscale.py

@@ -42,7 +42,7 @@ class OldGrayscale(BaseImageAugmentationLayer):
             after the `Grayscale` operation:
                  a. (..., height, width, 1) if output_channels = 1
                  b. (..., height, width, 3) if output_channels = 3.
-    Usage:
+    Example:
     ```python
     (images, labels), _ = keras.datasets.cifar10.load_data()
     to_grayscale = keras_cv.layers.preprocessing.Grayscale()

benchmarks/vectorized_jittered_resize.py

@@ -53,7 +53,7 @@ class OldJitteredResize(BaseImageAugmentationLayer):
     are translated and scaled according to the random scaling and random
     cropping.
 
-    Usage:
+    Example:
     ```python
     train_ds = load_object_detection_dataset()
     jittered_resize = layers.JitteredResize(

benchmarks/vectorized_mosaic.py

@@ -66,7 +66,7 @@ class OldMosaic(BaseImageAugmentationLayer):
         - [Yolov5 implementation](https://github.com/ultralytics/yolov5).
         - [YoloX implementation](https://github.com/Megvii-BaseDetection/YOLOX)
 
-    Sample usage:
+    Example:
     ```python
     (images, labels), _ = keras.datasets.cifar10.load_data()
     labels = tf.one_hot(labels,10)

benchmarks/vectorized_random_color_jitter.py

@@ -72,7 +72,7 @@ class OldRandomColorJitter(BaseImageAugmentationLayer):
             pass a tuple with two identical floats: `(0.5, 0.5)`.
         seed: Integer. Used to create a random seed.
 
-    Usage:
+    Example:
     ```python
     (images, labels), _ = keras.datasets.cifar10.load_data()
     color_jitter = keras_cv.layers.RandomColorJitter(

keras_cv/bounding_box/converters.py

@@ -338,7 +338,7 @@ def convert_format(
     performance loss, as each image will need to be processed separately due to
     the mismatching image shapes.
 
-    Usage:
+    Example:
 
     ```python
     boxes = load_coco_dataset()

keras_cv/bounding_box/to_ragged.py

@@ -29,7 +29,7 @@ def to_ragged(bounding_boxes, sentinel=-1, dtype=tf.float32):
     checking for padded sentinel value of the class_id axis of the
     bounding_boxes.
 
-    Usage:
+    Example:
     ```python
     bounding_boxes = {
         "boxes": tf.constant([[2, 3, 4, 5], [0, 1, 2, 3]]),

keras_cv/core/factor_sampler/constant_factor_sampler.py

@@ -29,7 +29,7 @@ class ConstantFactorSampler(FactorSampler):
     Args:
         value: the value to return from `__call__()`.
 
-    Usage:
+    Example:
     ```python
     constant_factor = keras_cv.ConstantFactorSampler(0.5)
     random_sharpness = keras_cv.layers.RandomSharpness(factor=constant_factor)

keras_cv/core/factor_sampler/normal_factor_sampler.py

@@ -31,7 +31,7 @@ class NormalFactorSampler(FactorSampler):
         min_value: values below min_value are clipped to min_value.
         max_value: values above max_value are clipped to max_value.
 
-    Usage:
+    Example:
     ```python
     factor = keras_cv.core.NormalFactor(
         mean=0.5,

keras_cv/core/factor_sampler/uniform_factor_sampler.py

@@ -31,7 +31,7 @@ class UniformFactorSampler(FactorSampler):
         seed: A shape int or Tensor, the seed to the random number generator.
             Must have dtype int32 or int64. (When using XLA, only int32 is
             allowed.)
-    Usage:
+    Example:
     ```python
     uniform_factor = keras_cv.UniformFactorSampler(0, 0.5)
     random_sharpness = keras_cv.layers.RandomSharpness(factor=uniform_factor)

keras_cv/datasets/imagenet/load.py

@@ -71,7 +71,7 @@ def load(
 ):
     """Loads the ImageNet dataset from TFRecords
 
-    Usage:
+    Example:
     ```python
     dataset, ds_info = keras_cv.datasets.imagenet.load(
         split="train", tfrecord_path="gs://my-bucket/imagenet-tfrecords"

keras_cv/datasets/pascal_voc/load.py

@@ -52,7 +52,7 @@ def load(
 ):
     """Loads the PascalVOC 2007 dataset.
 
-    Usage:
+    Example:
     ```python
     dataset, ds_info = keras_cv.datasets.pascal_voc.load(
         split="train", bounding_box_format="xywh", batch_size=9

keras_cv/keypoint/converters.py

@@ -82,7 +82,7 @@ def convert_format(keypoints, source, target, images=None, dtype=None):
     will need to be processed separately due to the mismatching image
     shapes.
 
-    Usage:
+    Example:
 
     ```python
     images, keypoints = load_my_dataset()

keras_cv/layers/augmenter.py

@@ -23,7 +23,7 @@ class Augmenter(keras.layers.Layer):
     Args:
         layers: A list of `keras.layers.Layers` to apply to the example
 
-    Examples:
+    Example:
     from keras_cv import layers
     images = np.ones((16, 256, 256, 3))
     augmenter = layers.Augmenter(

keras_cv/layers/feature_pyramid.py

@@ -68,7 +68,7 @@ class FeaturePyramid(keras.layers.Layer):
             levels. Defaults to None, and a `keras.Conv2D` layer with kernel 3x3
             will be created for each pyramid level.
 
-    Sample Usage:
+    Example:
     ```python
 
     inp = keras.layers.Input((384, 384, 3))

keras_cv/layers/fusedmbconv.py

@@ -72,7 +72,7 @@ class FusedMBConvBlock(keras.layers.Layer):
         block
 
 
-    Example usage:
+    Example:
 
     ```
     inputs = tf.random.normal(shape=(1, 64, 64, 32), dtype=tf.float32)

keras_cv/layers/hierarchical_transformer_encoder.py

@@ -51,7 +51,7 @@ class HierarchicalTransformerEncoder(keras.layers.Layer):
             `SegFormerMultiheadAttention`. If set to > 1, a `Conv2D`
              layer is used to reduce the length of the sequence. Defaults to `1`.
 
-    Basic usage:
+    Example:
 
     ```
     project_dim = 1024

keras_cv/layers/mbconv.py

@@ -85,7 +85,7 @@ def __init__(
             block
 
 
-        Example usage:
+        Example:
 
         ```
         inputs = tf.random.normal(shape=(1, 64, 64, 32), dtype=tf.float32)

keras_cv/layers/object_detection/anchor_generator.py

@@ -51,7 +51,7 @@ class AnchorGenerator(keras.layers.Layer):
       clip_boxes: whether to clip generated anchor boxes to the image
         size, defaults to `False`.
 
-    Usage:
+    Example:
     ```python
     strides = [8, 16, 32]
     scales = [1, 1.2599210498948732, 1.5874010519681994]

keras_cv/layers/object_detection/box_matcher.py

@@ -67,7 +67,7 @@ class BoxMatcher(keras.layers.Layer):
         ValueError: if `thresholds` not sorted or
         len(`match_values`) != len(`thresholds`) + 1
 
-    Usage:
+    Example:
 
     ```python
     box_matcher = keras_cv.layers.BoxMatcher([0.3, 0.7], [-1, 0, 1])

keras_cv/layers/object_detection/roi_generator.py

@@ -69,7 +69,7 @@ class ROIGenerator(keras.layers.Layer):
             feature maps / levels (as in FPN) this number is per
             feature map / level.
 
-    Usage:
+    Example:
     ```python
     roi_generator = ROIGenerator("xyxy")
     boxes = {2: tf.random.normal([32, 5, 4])}

keras_cv/layers/object_detection/roi_pool.py

@@ -40,7 +40,7 @@ class ROIPooler(keras.layers.Layer):
         image_shape: List of Tuple of 3 integers, or `TensorShape` of the input
             image shape.
 
-    Usage:
+    Example:
     ```python
     feature_map = tf.random.normal([2, 16, 16, 512])
     roi_pooler = ROIPooler(bounding_box_format="yxyx", target_size=[7, 7],

keras_cv/layers/overlapping_patching_embedding.py

@@ -39,7 +39,7 @@ def __init__(self, project_dim=32, patch_size=7, stride=4, **kwargs):
             stride: integer, the stride to use for the patching before
                 projection. Defaults to `5`.
 
-        Basic usage:
+        Example:
 
         ```
         project_dim = 1024

keras_cv/layers/preprocessing/aug_mix.py

@@ -60,7 +60,7 @@ class AugMix(BaseImageAugmentationLayer):
         - [Official Code](https://github.com/google-research/augmix)
         - [Unofficial TF Code](https://github.com/szacho/augmix-tf)
 
-    Sample Usage:
+    Example:
     ```python
     (images, labels), _ = keras.datasets.cifar10.load_data()
     augmix = keras_cv.layers.AugMix([0, 255])

keras_cv/layers/preprocessing/channel_shuffle.py

@@ -35,7 +35,7 @@ class ChannelShuffle(VectorizedBaseImageAugmentationLayer):
         groups: Number of groups to divide the input channels, defaults to 3.
         seed: Integer. Used to create a random seed.
 
-    Usage:
+    Example:
     ```python
     (images, labels), _ = keras.datasets.cifar10.load_data()
     channel_shuffle = ChannelShuffle(groups=3)

keras_cv/layers/preprocessing/equalization.py

@@ -35,7 +35,7 @@ class Equalization(VectorizedBaseImageAugmentationLayer):
         bins: Integer indicating the number of bins to use in histogram
             equalization. Should be in the range [0, 256].
 
-    Usage:
+    Example:
     ```python
     equalize = Equalization()
 

keras_cv/layers/preprocessing/fourier_mix.py

@@ -35,7 +35,7 @@ class FourierMix(BaseImageAugmentationLayer):
     References:
         - [FMix paper](https://arxiv.org/abs/2002.12047).
 
-    Sample usage:
+    Example:
     ```python
     (images, labels), _ = keras.datasets.cifar10.load_data()
     fourier_mix = keras_cv.layers.preprocessing.FourierMix(0.5)

keras_cv/layers/preprocessing/grayscale.py

@@ -41,7 +41,7 @@ class Grayscale(VectorizedBaseImageAugmentationLayer):
                  a. (..., height, width, 1) if output_channels = 1
                  b. (..., height, width, 3) if output_channels = 3.
 
-    Usage:
+    Example:
     ```python
     (images, labels), _ = keras.datasets.cifar10.load_data()
     to_grayscale = keras_cv.layers.preprocessing.Grayscale()

keras_cv/layers/preprocessing/grid_mask.py

@@ -82,7 +82,7 @@ class GridMask(BaseImageAugmentationLayer):
             [0 to 255]
         seed: Integer. Used to create a random seed.
 
-    Usage:
+    Example:
     ```python
     (images, labels), _ = keras.datasets.cifar10.load_data()
     random_gridmask = keras_cv.layers.preprocessing.GridMask()

keras_cv/layers/preprocessing/jittered_resize.py

@@ -68,7 +68,7 @@ class JitteredResize(VectorizedBaseImageAugmentationLayer):
             `"mitchellcubic"`.
         seed: (Optional) integer to use as the random seed.
 
-    Usage:
+    Example:
     ```python
     train_ds = load_object_detection_dataset()
     jittered_resize = layers.JitteredResize(

keras_cv/layers/preprocessing/mix_up.py

@@ -36,7 +36,7 @@ class MixUp(BaseImageAugmentationLayer):
         - [MixUp paper](https://arxiv.org/abs/1710.09412).
         - [MixUp for Object Detection paper](https://arxiv.org/pdf/1902.04103).
 
-    Sample usage:
+    Example:
     ```python
     (images, labels), _ = keras.datasets.cifar10.load_data()
     images, labels = images[:10], labels[:10]

keras_cv/layers/preprocessing/mosaic.py

@@ -72,7 +72,7 @@ class Mosaic(VectorizedBaseImageAugmentationLayer):
         - [Yolov5 implementation](https://github.com/ultralytics/yolov5).
         - [YoloX implementation](https://github.com/Megvii-BaseDetection/YOLOX)
 
-    Sample usage:
+    Example:
     ```python
     (images, labels), _ = keras.datasets.cifar10.load_data()
     labels = tf.one_hot(labels,10)

keras_cv/layers/preprocessing/posterization.py

@@ -37,7 +37,7 @@ class Posterization(BaseImageAugmentationLayer):
         bits: integer, the number of bits to keep for each channel. Must be a
             value between 1-8.
 
-    Usage:
+    Example:
     ```python
     (images, labels), _ = keras.datasets.cifar10.load_data()
     print(images[0, 0, 0])

keras_cv/layers/preprocessing/rand_augment.py

@@ -67,7 +67,7 @@ class RandAugment(RandomAugmentationPipeline):
         geometric: whether to include geometric augmentations. This
             should be set to False when performing object detection. Defaults to
             True.
-    Usage:
+    Example:
     ```python
     (x_test, y_test), _ = keras.datasets.cifar10.load_data()
     rand_augment = keras_cv.layers.RandAugment(

keras_cv/layers/preprocessing/random_apply.py

@@ -39,7 +39,7 @@ class RandomApply(BaseImageAugmentationLayer):
             but currently doesn't work with XLA. Defaults to False.
         seed: integer, controls random behaviour.
 
-    Example usage:
+    Example:
     ```
     # Let's declare an example layer that will set all image pixels to zero.
     zero_out = keras.layers.Lambda(lambda x: {"images": 0 * x["images"]})

keras_cv/layers/preprocessing/random_augmentation_pipeline.py

@@ -36,7 +36,7 @@ class RandomAugmentationPipeline(BaseImageAugmentationLayer):
     This layer can be used to create custom policies resembling `RandAugment` or
     `AutoAugment`.
 
-    Usage:
+    Example:
     ```python
     # construct a list of layers
     layers = keras_cv.layers.RandAugment.get_standard_policy(

keras_cv/layers/preprocessing/random_brightness.py

@@ -53,7 +53,7 @@ class RandomBrightness(VectorizedBaseImageAugmentationLayer):
       be clipped to the range `[0, 255]`, the valid range of RGB colors, and
       rescaled based on the `value_range` if needed.
 
-    Usage:
+    Example:
     ```python
     (images, labels), _ = keras.datasets.cifar10.load_data()
     random_brightness = keras_cv.layers.preprocessing.RandomBrightness()

keras_cv/layers/preprocessing/random_channel_shift.py

@@ -49,7 +49,7 @@ class RandomChannelShift(BaseImageAugmentationLayer):
             less channels.
         seed: Integer. Used to create a random seed.
 
-    Usage:
+    Example:
     ```python
     (images, labels), _ = keras.datasets.cifar10.load_data()
     rgb_shift = keras_cv.layers.RandomChannelShift(value_range=(0, 255),

keras_cv/layers/preprocessing/random_choice.py

@@ -28,7 +28,7 @@ class RandomChoice(BaseImageAugmentationLayer):
     `call`(), the policy selects a random layer from the provided list of
     `layers`. It then calls the `layer()` on the inputs.
 
-    Usage:
+    Example:
     ```python
     # construct a list of layers
     layers = keras_cv.layers.RandAugment.get_standard_policy(

keras_cv/layers/preprocessing/random_color_jitter.py

@@ -67,7 +67,7 @@ class RandomColorJitter(VectorizedBaseImageAugmentationLayer):
             pass a tuple with two identical floats: `(0.5, 0.5)`.
         seed: Integer. Used to create a random seed.
 
-    Usage:
+    Example:
     ```python
     (images, labels), _ = keras.datasets.cifar10.load_data()
     color_jitter = keras_cv.layers.RandomColorJitter(

keras_cv/layers/preprocessing/random_contrast.py

@@ -58,7 +58,7 @@ class RandomContrast(VectorizedBaseImageAugmentationLayer):
             `mean` is the mean value of the channel.
         seed: Integer. Used to create a random seed.
 
-    Usage:
+    Example:
     ```python
     (images, labels), _ = keras.datasets.cifar10.load_data()
     random_contrast = keras_cv.layers.preprocessing.RandomContrast()

keras_cv/layers/preprocessing/random_cutout.py

@@ -58,7 +58,7 @@ class RandomCutout(VectorizedBaseImageAugmentationLayer):
             when `fill_mode="constant"`.
         seed: Integer. Used to create a random seed.
 
-    Sample usage:
+    Example:
     ```python
     (images, labels), _ = keras.datasets.cifar10.load_data()
     random_cutout = keras_cv.layers.preprocessing.RandomCutout(0.5, 0.5)

keras_cv/layers/preprocessing/random_hue.py

@@ -48,7 +48,7 @@ class RandomHue(VectorizedBaseImageAugmentationLayer):
             preprocessing pipeline is set up.
         seed: Integer. Used to create a random seed.
 
-    Usage:
+    Example:
     ```python
     (images, labels), _ = keras.datasets.cifar10.load_data()
     random_hue = keras_cv.layers.preprocessing.RandomHue()

keras_cv/layers/preprocessing/random_jpeg_quality.py

@@ -35,7 +35,7 @@ class RandomJpegQuality(BaseImageAugmentationLayer):
         `tf.image.adjust_jpeg_quality()`.
         seed: Integer. Used to create a random seed.
 
-    Usage:
+    Example:
     ```python
     layer = keras_cv.RandomJpegQuality(factor=(75, 100)))
     (images, labels), _ = keras.datasets.cifar10.load_data()