A gist showing code for preprocessing images for tf.keras

tf_2_image_dataset.py

AUTOTUNE = tf.data.experimental.AUTOTUNE

path_ds = tf.data.Dataset.from_tensor_slices(train_file_list)
image_ds = path_ds.map(load_and_preprocess_image, num_parallel_calls=AUTOTUNE)
label_ds = tf.data.Dataset.from_tensor_slices(tf.cast(train_label_list, tf.int64))
image_label_ds = tf.data.Dataset.zip((image_ds, label_ds))

ds = image_label_ds.shuffle(buffer_size=1000 * BATCH_SIZE)
ds = ds.repeat()
ds = ds.batch(BATCH_SIZE)
# `prefetch` lets the dataset fetch batches, in the background while the model is training.
ds = ds.prefetch(buffer_size=AUTOTUNE)
ds

tf_2_preprocess_image.py

def preprocess_image(image):
  image = tf.image.decode_jpeg(image, channels=NUM_CHANNELS)
  image = tf.image.resize(image, [HEIGHT, WIDTH])
  image /= 255.0  # normalize to [0,1] range

  return image

def load_and_preprocess_image(path):
  image = tf.io.read_file(path)
  return preprocess_image(image)