Deploying Machine Learning into Docker Container

TASK -01

Task Description :-

👉 Pull the Docker container image of CentOS image from DockerHub and create a new container.

👉 Install the Python software on the top of docker container.

👉 In Container you need to copy/create machine learning model which you have created in jupyter notebook.

To perform this task we are using REDHAT LINUX(RHEL-8) as our Base OS for deploying machine learning model into the docker container and the dataset has been taken from the Internet.

Note: Kaggle is website where we can find many real-time datasets to work on.

STEPS TO COMPLETE THE TASK:

STEP - 1

Installation of Docker into RHEL-8

• # yum install docker -ce --nobest : is the command used to install docker in RHEL-8..

• To check if docker has been successfully installed use #docker info command

STEP -2

• Now we need to start the docker by using the command #systemctl start docker
• To check if our docker is running or not using the command #systemctl status docker

• By this we can see that Docker is actively running.

STEP-3

• Once the docker starts running, we need to pull the Docker container image of CentOS image from DockerHub and create a new container using the command #docker pull <image name:version>

• After this we need to run the docker container, the command used for it is  #docker run -it --name <> centos. In this we need to give a name of our choice, I gave 'Salary', a new OS opens up.

STEP-4

Installing the Python software on the top of docker container using the command #yum install python

STEP-5

Once the python is installed , we need to install 3 libraries

1. Pandas using #pip3 install pandas

2. scikit_learn using #pip3 install scikit_learn

3. numpy using #pip3 install numpy (numpy was already installed in my case)

After all these libraries have been installed , we need create a model in our container.

STEP-6

Now we need to create a directory in the container following the commands below ;

#mkdir /root/salary

#cd /root/salary

#ls

Once the directory has been created make sure to check if the dataset file has been saved in my case('Salary.csv') in the downloads directory in the main host OS(RHEL-8). Now copy that file location in the container using the command #docker cp / <source location> <container name/ID> <destination location>

This shows that the dataset has been stored properly

STEP-7

In this step we are going to train our model and to do so we need to provide a code to the OS by creating a file to type the code. The command used to ope the file is # vi <name>.py

A new text editor will be opened and we need to type our code in it using i:insert and to save and exit use Esc+:wq

To check if the file is running or not use #python3 <filename>.py

Once it runs properly ,we can test our model now.

STEP-8

To test the model we need to create one more file by using the command #vi<testingfilename>.py.

A new test editor will open up and we need to write the code shown.

After this we need to run this file using the same command  #python3<testingfilename>.py and it will display the output.

In this step we need to enter the year of experience to check the salary from, I have given the input as 3 and the output has been displayed. By we can predict the salary based on the year of experience using a simple dataset

Therefore by following the above steps, we can deploy a machine learning model into docker container.

Thankyou for reading my blog. Hope you like it!

P.s : This is my very first attempt of writing a blog , I would appreciate if you drop in your suggestions so that I can enhance my writing skills futher :)