how do you get the normalize value? I got array([128.05893, 127.94143, 128.40465]) for the mean and here is the code
hotdogs_train_sum = np.zeros(shape=(3, ))
hotdogs_test_sum = np.zeros(shape=(3, ))
for i in range(len(train_imgs)):
hotdogs_train_sum += train_imgs[i][0].sum(axis=[0, 1])
for j in range(len(test_imgs)):
hotdogs_test_sum += test_imgs[j][0].sum(axis=[0, 1])
(hotdogs_train_sum + hotdogs_test_sum) / (len(train_imgs) + len(test_imgs))