Google Colab Tutorial for Beginners | Get Started with Google Colab

INSTRUCTOR: Google
Colaboratory, or Colab for short

is the best way to get started
with artificial intelligence

and data science.

Everything you need all the
major libraries and tools

are already pre-installed for
you and pre-set up for you.

So by the time it
takes you to search

how to install
another software, you

can already be up and running
coding your data science

and artificial
intelligence code.

With Google Colab, you
can write, run, and share

Python code in your
browser, and this

allows you to get started
from your phone, tablet,

or laptop without
any configuration,

and you can immediately
share your code

with anyone with a simple link.

And everything you do in Colab
will be saved under your Google

account.

So you can access
them, share them,

just like the rest of your
files under your Google Drive.

If you are familiar with
the Jupyter Notebooks,

Google Colab is essentially
a Jupyter Notebook hosted

on Google servers with
additional functionalities

that makes it especially
popular among data science

artificial intelligence
communities.

So without further
ado, let's go ahead

and get started with Colab.

So first thing you want to do
is to search for Google Colab,

and it should be the one that
says colab.research.google.com.

Let's go into that.

So the very first thing you
should know about Google Colab

is that everything is
organized under notebooks.

So the first thing you see
when you come to the website

are basically your
recent notebooks.

And the first one you
will see will probably

be the Welcome to
Colaboratory notebook.

So to get started, first
thing we want to do

is to go ahead and
create a new notebook.

So the first time you
open a Colab notebook,

it will name it for
you with untitled 0.

In this case, it is 3 for me.

And one thing you want to do
is to rename this notebook

so that it's more
meaningful to us.

It can be something like
Colab Intro Tutorial.

Since many times, you will be
working with multiple notebooks

at a time, it can be really
helpful to appropriately name

your notebooks so if you
come back to it months

after writing it, it will be
descriptive and much easier

for you to know what you
were doing in that notebook.

And everything I
do in this notebook

will be saved to my Google
Drive under my Google account.

So when first we take a look at
Google Colab, the first thing

we will see, and the
major thing we will see

is something called a code
cell and a Play button.

We can go ahead and write
some Python code here

and it will be executed
not on our laptop,

but it will be run
on Google servers.

And currently, we are not
connected to that servers,

but we don't have to click
here to be connected.

But instead, usually
what you want to do

is to write some code here and
it will be automatically connect

to Google servers.

So let's write a
simple Python code.

Now, there are a couple of
ways I can run this code.

First one and the default one
is to click on this Play button.

And we just ran our
first Python code.

Usually, you'll want to
use keyboard shortcuts.

And the shortcut to
run a single cell

is Command Enter, which does
the same thing as clicking

on that Play button.

But most of the
time, chances are

you'll want to run a code
cell and create a new code

cell under it so
you can keep coding.

To do that, we will
use Shift Enter.

Which I just did.

Notice that there are two
types of cells I can create.

One of them is a code cell.

And this is where
we write our code.

Another one is a text cell.

So although it says
text cell, text

is not the only thing
we can write here.

Once we write some text,
using the icons above,

we can turn it
into bold, italic,

title, as well as
other features,

such as images or videos,
lists, and bullet points.

And there's actually even
a language for that called

markdown language.

And as you're getting
started with Colab,

it's not necessary to
learn markdown language,

but if you are
interested, you can

do that, especially later on
in your artificial intelligence

or data science journey.

For now, what you
should know is that we

can create code cells
that are modular,

and we can also use text cells
to make more explanations

or display images
that will be helpful

and maybe explain
our code better

with some rich text format.

So when we write code
inside code cells,

we said that it's not
running in our own laptop,

but it's running
at Google servers.

And there are actually
three types of servers

that Google can allocate to us,
and we can change it from here.

If we go to Runtime and
Change Runtime Type,

you will see that
you have the option

to have a hardware accelerator.

Right now, by
default, you are not

given a hardware accelerator,
but if you need one,

it's possible to have
either a GPU or a TPU.

So why would you need them?

If you're just getting started
and writing Python code

or simple machine learning
code that runs fast enough,

most likely, you'll be fine
with the default version, which

has no hardware
accelerator, meaning

that it will run on CPUs.

But Azure artificial
intelligence code

gets larger and more complex
as you do more operations.

That operations you make
will take more time,

and you will want to use some
kind of hardware accelerator.

And the most common hardware
accelerator is a GPU.

And we have access to
three GPUs within Colab.

One thing you will
notice is that when

you change the hardware
accelerator type from none

to GPU or a TPU, it
will basically tell you

that if you don't need
a hardware accelerator,

do not use a
hardware accelerator.

And this way, Google can
allocate its GPUs or TPUs

to people who need more
computational power.

For now, I'm just writing
simple Python code

so I can change it to none.

One thing you should
know is that when

you leave your notebook
unused for hours,

your notebook will be
disconnected from Google servers

to more efficiently
allocate its resources.

But do not worry.

When that happens,
all of your code

will be still safe and
nothing changed on them.

But you will need
to rerun your code.

And to do that, you can also
do it from Runtime and Run All.

So we talked about CPUs and
GPUs that you can connect to,

but there is also
another one which

we can access from Change
Runtime Type and Hardware

Accelerator, and that was TPU.

And TPUs are Tensor
Processing Units.

And do not be
scared by the name.

These are basically processing
units or processing chips

that are specifically designed
for significantly accelerating

artificial intelligence code.

And we also have free
access to them from Colab.

And this can be
especially useful

if you have larger
machine learning

models or artificial
intelligence models.

But if you are using
a smaller model,

you will be most likely
fine with the GPU

or non-hardware accelerator.

So until now, we just
seen Python code,

but we didn't see anything about
data science, machine learning,

or artificial intelligence code.

But one of the major advantages
of using Google Colab

is that you can't only code
Python code in the browser,

but all of the
major data science

and artificial intelligence
libraries and tools

come already
pre-installed in Colab.

And we can start using them
with import statements such as--

so in order to use TensorFlow,
NumPy, pandas, matplotlib,

and PyTorch, and
many more of them,

since they are already
pre-installed for us,

we can start using them with
these import statements.

And if you want, we can
also check their versions.

And as you can see, we're
using the latest versions

of the libraries, such as
using TensorFlow 2.4, which is

the latest one as of right now.

And if you're not familiar with
any of these, do not worry.

As you get started
with Colab, it

will be much easier to learn
them as you're coding them.

So as you're getting
started, you're not

required to know
what this code does

or what are these
libraries are about,

and you can learn them
one video at a time.

And every now and
then, you will want

to use libraries that
are not inside Colab.

And when you want
to do that, you

will want to prefix your code
with an exclamation mark.

And with that, you can install
other libraries to Colab

as well.

And when you want
to use them, you

can start using them with
import statements as well.

So as you are working
within a notebook,

if you want to create
a new notebook,

you can do that from
File and New Notebook.

And similar notebook actions,
such as opening the existing

ones, you can also do
them from here as well.

A lot of the times, many of the
machine learning, data science,

or artificial intelligence
codes are written in Colab,

and when you're learning from
them, one thing you want to do

is to save a copy in Drive.

And when you save
a copy in Drive,

you can have a copy
of that tutorial

and save it under your
Google Drive as well.

And if I do that, I
will have all the code

from that previous notebook,
except in this one,

I am not connected.

Which I can connect
in a second as well.

And I will have
this title change

to a copy of Colab Intro
Tutorial from Colab Intro

Tutorial.

And if I want to run
this code, you can either

run that code cell
by cell or you

can run that from
Runtime and Run

All, which will run all
the cells sequentially.

And that's one of
the reasons why

Colab is so popular among
data science and machine

learning communities.

And you can find
a lot of tutorials

are written in Colab for
ease of shareability,

because you can literally
share them with a single link,

and everyone that has an
internet connection and a Google

account can have
access to that code

and run them themselves,
as well as modifying it.

And if you want to
download the notebook,

you can also do that from here.

Under File, you can go
Download this notebook

as a ipynb file, which is
what we are using here.

And you can also download
it as a Python file.

One thing you should know
is that currently, we

are working with the ipynb
file, and this is the same file

format that you use
in Jupyter Notebooks.

And ipynb files are more popular
with data science and machine

learning because they have
additional capabilities on top

of Python files.

Some of those features
are code cells,

which you can save your
outputs of the code cells.

You can also change the
order of the code cells

more easily from this
up or down arrows.

You can delete
them, comment them,

and they're overall in
a more modular structure

that are a better fit for data
science and machine learning

code.

So these were everything
you needed to get

started with Google Colab.

As you just have seen, you can
just search for Google Colab

and get coding within
seconds without any setup.

And everything you do will be
saved under your Google account

and you can reach them
from your Google Drive

and share them
with a single link.

So this was it
for this tutorial.

If you learn better by
reading, by the way,

I also have this tutorial
in article format as well,

and I will make sure to put a
link to it in the description

box.

With that, I hope you got
some value out of this video.

If you want to see
more machine learning

and artificial
intelligence tutorials,

make sure to like and subscribe,
and I will see you next time.