hey my name is Felipe and welcome to my channel, in this
video I'm going to show you a very easy way to train an image classifier, this is going to be a
very easy and a very quick tutorial so let me show you. We are going to work with Google colab, this
is a notebook in my Google colab, and we're also going to work with Google Drive so please make
sure you create a directory in your Google Drive in order to work on this project, now let me show
you the data we are going to use today or actually let me show you the data I am going to use in
order to train this image classifier, I'm going to work with the fashion mnist dataset which is
a very popular dataset in computer vision you are going to find this data set a lot in different
computer vision related tutorials or actually different image classification related tutorials,
this is a very popular dataset and this dataset is comprised with many many different fashion
related items right, you can see these are a few samples these are a few images of this
dataset and all of these images are fashion related items, let me show you all the different
categories we have in this dataset these are the 10 different categories we have and you can see
these categories are t-shirt trouser pullover dress coat sandal shirt sneaker bag and ankle
boot so once we train an image classifier with this dataset we will be able to classify an image
into one of these 10 categories, so I'm going to post a link to this dataset in the description
of this video so if you want to follow along this tutorial you can just go ahead and download
exactly the same dataset I am going to use today, now let me show you something which is very
important and it's how you need to structure the data in order to work on this tutorial, so now let
me show you how to structure the data this is a very important step so let me show you how to do
it, you can see I have two directories one of them is called train and the other one is called test,
it's very important you name these directories exactly like this one of them should be called
train and the other one should be called test, I know this sounds like a detail it sounds like
it's a completely meaningless and unimportant detail but trust me it's very important you name
these directories exactly like this, absolutely every single technology you use in order to train an
image classifier or any type of machine learning model is going to have its own requirements
regarding how to structure the data and in this tutorial we are going to use ultrallitics version
of YOLOV8 so this is how we need to structure the data in order to use YOLOV8, it's very important
you create two directories and it's very important you call these directories exactly like this one
of them should be called train and the other one should be called test... you could also call this
or directory differently you could call this directory val, if you call this directory val is
because this is going to be the validation data if you call the directory test this is the test data
but in order to keep things simple I invite you to just call these directories exactly like this
one of the them is called train the other one is called test if you're going to follow along this
tutorial then just create these two directories and name these two directories exactly like this and
everything is going to work exactly the same as it's going to work for me, so now once you have
created these two directories you can see that within train and within test I have 10 folders and
these 10 folders are called, are named, exactly the same as the categories we are going to classify
with this classifier remember I am going to use the fashion mnist dataset and I'm going to
classify all my images into one of these 10 categories I have over here so this is exactly
why I have 10 folders within train and within test and these folders are called exactly the same
as the categories I am going to classify all my images into and let me show you a few examples now
if I open the sneaker folder you can see that this is where I have all my images of the sneaker
category, you can see that this is the folder where I have all the images of sneakers and if I
go up and I show you another example for example t-shirt you can see that this is where I have
all my t-shirts and so on right so each one of these folders has all the images of that specific
category so in order to structure the data into the exact structure you need in order to train
this image classifier using YOLO V8 you need two directories one of them is called train the one
is called test then within train and within test you need as many folders as categories
you are going to classify and then within each one of these folders is where you are going to locate
all of your data, this is all of your training data and this is all of your test data, so once you
structure the data exactly like this the only thing you need to do is to zip this directory
where you have located your train directory and your test directory and then then the only thing
you need to do is to compress this directory and to create this file I have over here which is data.zip
and then once you have this file the only thing you need to do is to get back to your Google Drive
and you need to upload this file into your Google Drive, in my case you can see that I have already
uploaded this file it's here data.zip so I'm not going to upload this file again but please make
sure to upload this file into your Google drive into the directory you have created in your Google
Drive in order to continue now let's get back to Google colab, let's get back to this notebook
in my Google collab and let me show you how to train this image classifier and remember this
notebook will be available in the GitHub repository of today's tutorial so just take a look at the
description of this video and you are going to find a link to the GitHub repository of today's
tutorial and this is where you are going to find this notebook so the only thing we need to do now
is to execute all of the cells in these notebooks one at the time and that's going to be pretty much
all in order to train this image classifier so I'm going to execute this cell over here this is going
to take care of mounting our Google drive into this Google colab and this is going to be pretty much
all we need to do in order to access the data in our Google Drive, and this is going to be pretty
much all, now remember we have uploaded all of our data over here in this file which is data.zip
we are going to copy this file which is in our Google Drive we're going to copy this file into
this Google collab environment into this Google colab instance, and then we are going to unzip all
the content of this file here right so it's very important you change this value over here which
is the location of the data you have uploaded into your Google Drive, you can see that now it
says train yolo8 image classification google colab data.zip which is the the location of the data in
my Google Drive right you can see this is train yolov8 image classification Google colab and the file is
called data.zip and everything it's in my root directory in my Google Drive which is called my
drive which is exactly what we have over here, so just make sure you change this value to whatever
location you have chosen for this file in your Google Drive and once you have changed this value
over here the only thing you need to do is to press enter and this going to copy this file into
your Google colab environment and then it's just going to unzip all the content
of this file and that's going to be pretty much all, okay you can see that we have extracted all
the content of this file into our Google colab right you can see that these are all the images
into in our dataset and that's going to be pretty much all so the next step in this process will be
executing this cell over here which is going to install ultralytics which is the Python package
we are going to use in order to train this image classifier and then the only thing we need to
do is to execute this cell over here you can see that we are importing ultralytics which is
the python package we just installed and then we are defining a new variable which is called
model and then we're just calling model.train and that's pretty much all we need to do in order to
train this image classifier and please notice that the data location is content data which is exactly
where we have copied all of our data over here right we have extracted this zp file exactly here
so this is why we are specifying this as our data location and then the number of epochs you can see I
have set the number of epochs in five so I'm going to train this image classifier for only five epochs but
you can train this classifier for as many many epochs as you want but in my case I think five epochs
are going to be enough in order to show you how to train this image classifier so you can see
that this is the training process everything is going just fine everything is just okay and now
the only thing we need to do is to wait until this is completed and once your training process
is completed the only thing you need to do is to get all the results from your training process
right you want to get the results you want to analyze your model, you want to get the weights of the
model you trained and this is where all the results were saved, in content runs, so the only thing you
need to do now is to execute this cell over here and this is going to take care of copying all the
content in this directory into your Google Drive and this going to be pretty much all in order to
get all these results right and please remember to change this value over here to the location
of your directory in your Google Drive and once you have changed this value over here the
only thing you need to do is to press enter and it's going to take care of copying all the files
into your Google Drive now let me show you if I go back to my Google Drive you can see that this
is the directory we have just copied over here which is called runs and this is where all the
results were saved so this is going to be pretty much all in order to show you how to train this
imagifier in order to show you a very very quick and a very easy way to train this image classifier
but if you want to have a much more comprehensive description of how to train this image classifier, if
you want much more details then I invite you to take a look at other of my previous videos where
I show you exactly the same process, I show you exactly how to train an image classifier using
YOLO V8, exactly the same process, but in a much more comprehensive way and I show you much more
details and I also show you how to make inferences with the model you trained, I show you how to analyze how
to evaluate all your training process, I show you a much more comprehensive description of the entire
process so I definitely invite you to take a look at that other video over there but for now this is
going to be all so thank you so much for watching this tutorial my name is Felipe, I a computer
vision engineer and see you on my next video.
