My old Tensorflow 2.8.0 code used ImageDataGenerator to load and augment my dataset. One of these data augmentation transformations was a random shear. However ImageDataGenerator is deprecated as of 2.13. Another option was to use the KerasCV library via tensorflow addons but again this is deprecated. tf.keras.preprocessing.image has a random shear function but it is deprecated. tf.image does not even have a random shear function. So my question is what is the acceptable way to implement random shear in Tensorflow 2.13?
Hi @Anon74881, In the pre processing function you can consider using keras_cv.layers.RandomShear. Thank You.