File size: 802 Bytes
5fa1a76 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 |
from tensorflow import keras
from tensorflow.keras import layers
size = (image_processor.size["height"], image_processor.size["width"])
train_data_augmentation = keras.Sequential(
[
layers.RandomCrop(size[0], size[1]),
layers.Rescaling(scale=1.0 / 127.5, offset=-1),
layers.RandomFlip("horizontal"),
layers.RandomRotation(factor=0.02),
layers.RandomZoom(height_factor=0.2, width_factor=0.2),
],
name="train_data_augmentation",
)
val_data_augmentation = keras.Sequential(
[
layers.CenterCrop(size[0], size[1]),
layers.Rescaling(scale=1.0 / 127.5, offset=-1),
],
name="val_data_augmentation",
)
Next, create functions to apply appropriate transformations to a batch of images, instead of one image at a time. |