» KerasCV
KerasCV
KerasCV is a toolbox of modular building blocks (layers, metrics, losses, data augmentation) that computer vision engineers can leverage to quickly assemble production-grade, state-of-the-art training and inference pipelines for common use cases such as image classification, object detection, image segmentation, image data augmentation, etc.
Quick links
Installation
KerasCV requires Python 3.7+ and TensorFlow 2.9+.
Install the latest release:
pip install keras-cv --upgrade
You can also check out other versions in our GitHub repository.
Quick Introduction
Create a preprocessing pipeline:
import keras_cv
import tensorflow as tf
from tensorflow import keras
import tensorflow_datasets as tfds
preprocessing_layers = [
keras_cv.layers.RandomResizedCrop(
target_size=(224, 224),
crop_area_factor=(0.8, 1.0),
aspect_ratio_factor=(3/4, 4/3)
),
keras_cv.layers.RandomFlip(),
keras_cv.layers.RandAugment(value_range=(0, 255)),
keras_cv.layers.CutMix(),
keras_cv.layers.MixUp()
]
def augment_data(images, labels):
labels = tf.one_hot(labels, 3)
inputs = {"images": images, "labels": labels}
for layer in preprocessing_layers:
inputs = layer(inputs)
return inputs['images'], inputs['labels']
Augment a tf.data.Dataset:
dataset = tfds.load('rock_paper_scissors', as_supervised=True, split='train')
dataset = dataset.batch(64)
dataset = dataset.map(augment_data)
Create a model:
densenet = keras_cv.models.DenseNet121(
include_rescaling=True,
include_top=True,
classes=3
)
densenet.compile(
loss='categorical_crossentropy',
optimizer='adam',
metrics=['accuracy']
)
Train your model:
densenet.fit(dataset)
Citing KerasCV
If KerasCV helps your research, we appreciate your citations. Here is the BibTeX entry:
@misc{wood2022kerascv,
title={KerasCV},
author={Wood, Luke and Zhu, Scott and Chollet, Fran\c{c}ois and others},
year={2022},
howpublished={\url{https://github.com/keras-team/keras-cv}},
}