Deploying ML models To the Web with Flask on AWS EC2 Instance

Creating a Machine Learning Model is not enough!. Deploying it in a web/android application will fulfill the real-world application. To create a web application we can use Django or Flask as a backend framework. Flask is a micro backend framework that is perfectly suitable for creating a web application with Machine Learning Models. After creating a web application we need to deploy it on the cloud to make it available for all. For deploying easily on the cloud we can use AWS EC2 Instance. So here are the following steps from scratch to deploy a flask application with ML models on AWS EC2 Instance.

Step by Step Process to Deploy a Flask Application with ML Models on AWS EC2 Instance:-

1. Creating a Machine Learning Model:-

Let’s create a basic machine learning model with IRIS Dataset to classify the different iris species. The features are sepal length (cm), sepal width (cm), petal length (cm), petal width (cm), and labels are types of iris species i.e Setosa, Versicolor, Virginica. Firstly we need to load the data from default sklearn datasets and split the features and labels into training and testing sets. As it is a classification problem we can use logistic regression and train it with the training set and predict the labels with the testing set for predicting the accuracy score.

The accuracy score for the above model results in 100% which is awesome!

2. Pickling the Model Object:-

We created the model and know we need to store the model object into a pickle file by dumping the object using the pickle library as shown below.

A file named model.pkl file is created which is around 956 bytes.

3. Creating a Website Template:-

Now we need to create a website template with a form consisting of 4 input fields and a predict button. A result field should also be assigned to display the predicted result. It should be named as index.html and need to be placed in the templates folder. All the other files like images, CSS files, JS files should be separately placed in a folder named static.

4. Creating an AJAX Request:-

Now once the form is created then we need to create an ajax request to URL ‘/predict’ bypassing the all form input values and need to get the response which is to be assigned to the result field as text. The following script needs to be added to the index.js or can be kept as the JS file which is to be stored in a static folder and to be linked to index.html.

5. Create a Flask Backend API:-

Now we need to create a python file for collecting the ajax request and to predict the type of iris species. For that firstly you need to install flask using pip in CMD.

pip install flask

We need to create a python file to process the ajax request. Do carefully see the comments in the below file to understand the code.

Once if the above-mentioned app.py is coded then we need to run the flask web application using the following command in CMD.

python app.py

When the above command is executed then the server is started locally at the URL http://127.0.0.1:5000/. In this way, we can deploy the flask application locally on the machine. Now our goal is to run it on AWS EC2 Instance.

Whenever we need to deploy our application on AWS, we need to change the host and port in which the following flask application is running. By default, the host address will be at ‘127.0.0.1’ in other words localhost and port number by default is 5000. Now we need to change the default port number to 8080 and host address to ‘0.0.0.0’. For this, the following code needs to be added in place of app.run().

Now if you run the flask application it will run but if you go to http://0.0.0.0:8080/ it will give you the error. The reason for this is address 0.0. 0.0 is a non-routable meta-address used to designate an invalid, unknown, or non-applicable target. Don’t care about this error just leave about this and go on to the next step.

6. Getting Started with AWS:-

Now it’s all set, the remaining thing to reach our goal is to deploy flask application to AWS Cloud. For this, we need to have an AWS account. If you are new to AWS, You need to create an AWS account by providing your credit card details to avail of free tier services free for one year. After creating the account you need to go to AWS Management Console and need to sign in to AWS by using the account that you have just created. Once you sign in the webpage looks similar to the below image.

AWS Management Console

The above-shown snapshot may vary by time as the site changes from time to time.

7. Creating an AWS EC2 Instance:-

Now we need to go to the search bar in the AWS Management Console and search for EC2 and need to click on running instances. Click on Launch Instance to create the new EC2 instance. Now we need to follow 4 steps to create an instance.

I. Choose an Amazon Machine Image (AMI):-

We need to choose AMI which means choosing an OS for the server. In this case, we can choose the UBUNTU 18.04 which is a free tier and proceed next.

Choosing AMI as UBUNTU Server

II. Choose Instance Type:-

Now we need to choose instance type for our server. In this case, we can go with a free tier instance. If yours is a company you can choose the paid ones. Here is the snapshot of the choose instance type.

Choosing Instance Type

Once you selected the instance we need to click on “Review and Launch”.

III.Edit Security Groups:-

Once the above steps are done, you need to click on the “Edit Security Groups” in the “Review and Launch” page. Once if you go “Edit Security Groups” Page you need to create a security group.

What is the reason behind creating Security Groups?

It helps us to edit which kind of traffic that should be allowed while accessing the instance and it helps us to decide which IPs can access our application which allows us to restrict the usage.

In this case, we are allowing all traffic and source to anywhere so that every network from anywhere in the world can be able to access our application. Once it is done we need to click on “Review and Launch”. Here is the Snapshot of editing security groups.

Editing Security Groups

IV. Creating a Private Key Pair:-

Once you followed with the above steps you will be able to click on the launch button in “Review Instance Launch”. You will be getting a pop-up for choosing a private key pair. Let’s click on create a new private key pair and choose a keypair name. In this case, I am choosing it as the iris. Now we need to click on the download key pair button to download it.

An Important Note is mentioned here below:-

Important Note

Now store this file in your project folder and click on launch instances. Here is the snapshot of how to create a private key pair.

how to create a private key pair

In this way, we can be able to create an EC2 Instance. Now we need to go to view instances where you could be able to find the instance you have created.

8. Downloading Required Softwares:-

The Required Softwares to be downloaded which helps in connecting the local machine with the AWS EC2 instance. The list of Softwares to do be downloaded are

  1. WinSCP:- It is a free and open-source SFTP, FTP, WebDAV, Amazon S3, and SCP client for Microsoft Windows. Its main function is secure file transfer between a local and a remote computer.
  2. Putty:- It is a free and open-source terminal emulator, serial console, and network file transfer application. It supports several network protocols, including SCP, SSH, Telnet, rlogin, and raw socket connection. It can also connect to a serial port.

9. Converting private key from PPM to PPK Format using PuttyGen:-

You need to convert a private key from PPM format which you have downloaded previously to PPK format using PuttyGen.It will be very much useful in the future for getting access to transfer files from local machine to Ubuntu Server. To convert firstly you need to open PuttyGen Software which is already installed default when you installed Putty. The Snapshot of PuttyGen looks like shown below.

Now you need to click on load and add the PPM file you recently downloaded. Then you need to click on “Save private key” and store it. Now close the PuttyGen window and continue to follow the next step.

11. Transferring Project files using WinSCP:-

Now you need to transfer project files from local machine to ubuntu server using WinSCP Software. Firstly you need to go to the AWS Management Console and right-click on the instance where you will find many options. Now click on connect and you will be able to get the similar page shown below:-

Now you need to copy the Public DNS which is shown in point no 4.Then you need to open WinSCP and need to paste this DNS address in the host input value. The window looks similar to below shown one.

Now click on advanced options and go to “Authentication” and insert the created private key in PPK format. The window looks similar to below shown one.

Now click on OK and go to log in. Now you can be able to connect with the ubuntu server. Now select the project files you wanted to add to the server and drag and drop to the right side section in the window. The file transfer starts and after a while files will be transferred. The window looks similar to below shown one.

12. Connect to the Instance using Git Bash:-

Now required file transfer is done and the remaining thing is we need to get an access terminal to compute the files. For that open git bash and navigate to the project source. Once you have done with that you need to go to instances in AWS Management Console click on connect then you will be able to find a similar page as shown below.

Now you need to go to the example section and copy the code given there. Now paste the command on git bash and click enter and you will be getting a question to accept the security. Enter “yes” and click on enter. Now you are connected to the instance from your local machine. The git bash looks similar to this when you successfully connected to the server instance.

13. Install the Required Libraries and run the Flask Application:-

Now once the terminal is ready now we need to run some commands before running the flask app. Firstly we need to update the OS and update pip to install libraries in the future. The command for performing above process is

sudo apt-get update && sudo apt-get install python3-pip

Now you need to install all required libraries to run the flask application. The required commands for this application are as follows:-

pip3 install numpy
pip3 install flask
pip3 install sklearn

Now all the libraries are installed now we need to run the application by entering following command on git bash.

python3 app.py

Now the application started running. You can view the application at the URL as <your DNS>::<your port no>.In my case it is http://ec2-3-20-236-91.us-east-2.compute.amazonaws.com:8080/

The web page is shown below:-

Important Note:-

Now your application is hosted but there is a problem you will face once you close your git bash. The problem is once if you close git bash and you reload the URL your screen looks similar to below.

How to Solve this problem?

To solve this issue we need to use Screen. Screen or GNU Screen is a terminal multiplexer. In other words, it means that you can start a screen session and then open any number of windows (virtual terminals) inside that session. Processes running in Screen will continue to run when their window is not visible even if you get disconnected.

For this, we need to run the following command on the git bash.

screen -R deploy python3 app.py

This helps your server to make active by creating multiple screens and makes it connected and run the server even if you close the git bash.

Final Deployed Flask Application with ML Model

Pro Tip:-

Once you don’t have a requirement to stop your server as it helps you to pay for how much you use.

Conclusion:-

In this way, you can be able to deploy a flask web application with ML models using AWS EC2 instances.

Thanks for reading my article. If you like my article do 👏👏 .

ShapeAI

LEARN | CODE | COMPETE | INTERN

ShapeAI

We at ShapeAI, are a team with an aim to provide students with a platform to research and make projects in the field of AI and ML. We aim to provide an interactive learning platform for students where they will be able to learn and develop projects with our support.

Sai Durga Kamesh Kota

Written by

Google DSC ML Lead | AI Practitioner🤖 | 3️⃣ rd year undergrad 🎓 in CSE 💻 at NIT Patna | Tech Blogger📝

ShapeAI

We at ShapeAI, are a team with an aim to provide students with a platform to research and make projects in the field of AI and ML. We aim to provide an interactive learning platform for students where they will be able to learn and develop projects with our support.

Medium is an open platform where 170 million readers come to find insightful and dynamic thinking. Here, expert and undiscovered voices alike dive into the heart of any topic and bring new ideas to the surface. Learn more

Follow the writers, publications, and topics that matter to you, and you’ll see them on your homepage and in your inbox. Explore

If you have a story to tell, knowledge to share, or a perspective to offer — welcome home. It’s easy and free to post your thinking on any topic. Write on Medium

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store