Learning rate for NN training

[pamio]

in tensorflow or torch, I have observed that the learning rate is usually one number like 0.001 or 0.01 etc. But in this tutorial I find that the learning rate for 16 classes is something like this

cntk_lr_per_image = [0.01] * 10 + [0.001] * 5 + [0.0001] which is array of 16 items which is equal to the number of classes. So if I have 80 classes to train should this array cntk_lr_per_image contain 80 learning rate ?

Something like this cntk_lr_per_image = [0.01] * 40 + [0.001] * 30 + [0.0001]*10 ?
I understand what learning rate is and how it impacts the overall learning of the model and there's no straight forward way to choose a learning rate but this tutorial uses an array of learning rates which is what confuses me.

Can you help me with understanding how learning rate us used in this example ?

[PatrickBue]

Hi, this is just a standard learning rate schedule used in most research work, where the learning rate is divided by 10 after a certain number of epochs