Transfer Learning with InceptionV3 model on Cats vs Dogs

#!/usr/bin/python3

import tensorflow as tf

class MyCB(tf.keras.callbacks.Callback):
	def on_epoch_end(self, epoch, logs={}):
		if (logs.get('loss') < 0.1):
			self.model.stop_training = True
			print("\nStopping Training..........\n")

cbs = MyCB()

dg = tf.keras.preprocessing.image.ImageDataGenerator(
	rescale=1/255.,
	rotation_range=40,
	width_shift_range=0.2,
	height_shift_range=0.2,
	shear_range=0.2,
	zoom_range=0.2,
	horizontal_flip=True
)
gen = dg.flow_from_directory(
	'data/training',
	target_size=(150,150),
	batch_size=20,
	class_mode='binary')

vdg = tf.keras.preprocessing.image.ImageDataGenerator(
	rescale=1/255.
)
vgen = vdg.flow_from_directory(
	'data/validation',
	target_size=(150,150),
	batch_size=20,
	class_mode='binary')

saved_model_weights='data/inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5'

ptm = tf.keras.applications.inception_v3.InceptionV3(
	input_shape=(150,150,3),
	include_top=False,
	weights=None)
ptm.load_weights(saved_model_weights)
for l in ptm.layers:
	l.trainable = False
last_output = ptm.get_layer('mixed7').output

x = tf.keras.layers.Flatten()(last_output)
x = tf.keras.layers.Dense(units=512, activation='relu')(x)
x = tf.keras.layers.Dropout(0.2)(x)
x = tf.keras.layers.Dense(units=1, activation='sigmoid')(x)

model = tf.keras.Model(ptm.input, x)

model.compile(
	optimizer=tf.keras.optimizers.RMSprop(learning_rate=0.0001),
	loss=tf.keras.losses.BinaryCrossentropy(),
	metrics=['accuracy'])

model.fit(
	gen,
	validation_data=vgen,
	epochs=20,
	callbacks=[cbs])