<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Anvita Joshi on Medium]]></title>
        <description><![CDATA[Stories by Anvita Joshi on Medium]]></description>
        <link>https://medium.com/@anvitajoshi7?source=rss-4ac8149925f8------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*sN_H2ijbZTR6b4-0f09htw.jpeg</url>
            <title>Stories by Anvita Joshi on Medium</title>
            <link>https://medium.com/@anvitajoshi7?source=rss-4ac8149925f8------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 08:52:35 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@anvitajoshi7/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Introduction to basic AWS Services]]></title>
            <link>https://medium.com/@anvitajoshi7/introduction-to-basic-aws-services-c862c657807d?source=rss-4ac8149925f8------2</link>
            <guid isPermaLink="false">https://medium.com/p/c862c657807d</guid>
            <dc:creator><![CDATA[Anvita Joshi]]></dc:creator>
            <pubDate>Mon, 31 Oct 2022 12:05:07 GMT</pubDate>
            <atom:updated>2022-10-31T12:05:07.218Z</atom:updated>
            <content:encoded><![CDATA[<p>Contents</p><ul><li>AWS Services introduction &amp; Networking basics</li><li>Compute — EC2, ELB, Elastic Beanstalk, Lightsail, Compute for Containers</li><li>Security — IAM</li><li>Storage — EBS, EFS, S3</li><li>Database — Amazon RDS and Dynamo DB</li><li>Data Integration — AWS Glue and Athena</li><li>Serverless Architecture — AWS Lambda</li><li>CDN &amp; Monitoring</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*b8gA5ZVyoYK3OOjP.jpg" /><figcaption>AWS Services</figcaption></figure><h3>1. AWS Services introduction &amp; Networking basics</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*spmb54k8tByY3JY2VBDbUw.png" /><figcaption>AWS basic 3-tier architecture</figcaption></figure><p><strong>AWS</strong> maintains multiple geographic <strong>Regions. </strong>Each region will have multiple Availability zones. Amazon Virtual Private Cloud (Amazon VPC) gives you full control over your virtual networking environment, including resource placement, connectivity, and security.</p><p>Once resources are placed in VPC, we can define how VPCs communicate with each other across accounts, Availability Zones, or AWS Regions. It allows to create Subnets, Route Tables, Elastic IPs, Site to Site VPN etc.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/670/1*kzTXaycj9_VYpW-vz8ROrw.png" /><figcaption>AWS Primary Network components</figcaption></figure><h3>2. Compute</h3><h4>i) Elastic Compute Cloud (EC2)</h4><p>Elastic Compute Cloud (EC2) are configurable virtual servers on AWS</p><p>AWS EC2 allows users to rent virtual computers on which they can run their own applications. The OS image, CPU Architecture, number of CPU cores, memory size, storage and networking is fully configurable.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/752/1*xH3YeoWigiiAXMpCWh-NRA.png" /><figcaption>Setting up EC2 instance</figcaption></figure><h4>ii) Elastic Load Balancer (ELB)</h4><p>ELB is used to distribute network traffic to improve application scalability</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/704/1*roclB7F4077UlwJbJc3JEQ.png" /><figcaption>ELB</figcaption></figure><p>Elastic Load Balancing (ELB) automatically distributes incoming application traffic across multiple targets and virtual appliances.</p><h4>iii) Elastic Beanstalk</h4><p>AWS Elastic Beanstalk is an easy-to-use service for deploying and scaling web applications and services developed with Java, .NET, PHP, Node.js, Python, Ruby, Go, and Docker on familiar servers such as Apache, Nginx, Passenger, and IIS.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/846/1*VaibTb3MDjGui62kiRKFAw.png" /><figcaption>Setting up Elastic Beanstalk</figcaption></figure><p>You can simply upload your code and Elastic Beanstalk automatically handles the deployment, from capacity provisioning, load balancing, auto-scaling to application health monitoring. At the same time, you retain full control over the AWS resources powering your application and can access the underlying resources at any time. Here, the resources created by Beanstalk such as EC2 instances are fully managed by AWS, therefore we need not manage things such as OS updates manually.</p><h4>iv) Amazon Lightsail</h4><p>Amazon Lightsail is <strong>a virtual private server (VPS) provider</strong> and is the easiest way to get started with AWS for developers, small businesses, students, and other users who need a solution to build and host their applications on cloud.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/818/1*JxV2CvVCSSVEvQTMFAlyzg.png" /><figcaption>Configuring a server on Lightsail</figcaption></figure><p>Lightsail is an integrated product of services offered by AWS. It is better for small to medium scale workloads</p><h4>v) Compute for Containers</h4><p>We have Amazon ECS (Elastic Container Service), EKS (Elastic Kubernetes Service) and AWS Fargate available for container orchestration</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AlR7Xpa5P2Fg0sdCunyM-Q.png" /><figcaption>Different container orchestration services available</figcaption></figure><p>Amazon <strong>Elastic Container Service</strong> (<strong>ECS)</strong> is a fully managed container orchestration service that helps you easily deploy, manage, and scale containerized applications.</p><p>Amazon <strong>Elastic Kubernetes Service</strong> (<strong>EKS</strong>) is a managed service and certified Kubernetes conformant to run Kubernetes on AWS and on-premises.</p><p>AWS <strong>Fargate</strong> is a serverless, pay-as-you-go compute engine that lets you focus on running containers without managing servers</p><h3>3. Security</h3><p><strong>IAM — Identity and Access Management</strong></p><p>IAM service of AWS provides Resource Authentication and Authorization</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/742/1*h7G9t0uvd2IOasRdj3M8yA.png" /><figcaption>IAM dashboard</figcaption></figure><p>IAM allows to create Users, Groups, Roles and Policies to control who has access to what resources on AWS. It also provides Security recommendations to keep the AWS account secure.</p><h3>4. Storage</h3><p>When it comes to storage on AWS there are 3 kinds of primary storage systems first is the <strong>Elastic Block Store</strong>, the <strong>Elastic File System</strong> and third is S<strong>imple Storage Service</strong> i.e S3</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*K7GaoiqwfnaS5UkRZBBcYQ.png" /></figure><h4>i) Elastic Block Store (EBS)</h4><p>EBS is block storage service used for EC2 compute instances, they deal with bulk data stored in HDDs or SSDs.</p><p>They are like C Drives or D drives in Windows systems and Linux based systems that is mounted on top of EC2 instances</p><p>These are volumes that can be partitioned and formatted</p><h4><strong>ii) Elastic File System (EFS)</strong></h4><p>EFS on the other hand mounts file systems to EC2 instances using NFS protocol</p><p>These cannot be partitioned and formatted unlike EBS systems</p><p>&amp; they can be used only with Linux based EC2 instances</p><h4><strong>iii) Simple Storage Service (S3)</strong></h4><p>S3 on the other hand is an Object Oriented storage service on which we can store any type of data, be it file, photo, video. These objects are grouped together into buckets that is stored on S3</p><p>S3 is a universal namespace, meaning each S3 bucket you create must have a unique name that is not being used by anyone else in the world</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/698/1*hhHT-BeSa1IJkG2w7MFViQ.png" /><figcaption>S3 Console — buckets</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/698/1*7yxHGk-3HONKPvIBsjZY7Q.png" /><figcaption>S3 Console — objects within a bucket</figcaption></figure><h3>4. Database</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/574/1*HteDEFCtRdnTWZbHn6YjCA.png" /><figcaption>Different types of DB services offered by AWS</figcaption></figure><p><strong>Amazon RDS</strong> stands for relational database service, as the name suggests this is used in case of relational databases where data is stored in structural format using rows and columns<br> <br> Next is <strong>Dynamo DB</strong> service, this is predominantly used for noSQL or <strong>non-tabular</strong> type of databases<br> <br><strong>Amazon Aurora (Aurora)</strong> is <strong>a </strong>fully managed relational database engine that’s compatible with MySQL and PostgreSQL</p><p><strong>Neptune</strong> is a DB service used for graph type of data</p><h4><strong>i) Amazon RDS</strong></h4><p>We can create a RDS database instance and connect to it remotely with the endpoint mentioned in the DB details and use it like a normal mySQL engine</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ec_e1uXn2dWpOZjNkxLC3g.png" /><figcaption>RDS dashboard on AWS Console</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/822/1*LtBUhMqR7zfLli-8zyt4hw.png" /><figcaption>Remote connection to RDS database instance</figcaption></figure><h4><strong>ii) Dynamo DB</strong></h4><p>With Dynamo DB, we can create non-structural dB like the ones shown in the picture on the right, we just give the table name &amp; set our partition key &amp; sort key which as primary &amp; secondary keys</p><p>Then we can create items under this as shown in the pictures below, we can create it using jsons or add values explicitly in its Form. We can add any number of additional attributes</p><p>Every attribute can be of different datatype here since its not constrained structural</p><p>Again this is a fully managed service from AWS, so we do not need to worry about the configuration of underlying resources</p><p>It is similar to mongoDB, has faster retrieval and is the cheapest noSQL solution available</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/432/1*FAvInrBvvKCWWzPJBXZqMQ.png" /><figcaption>Creation of a Dynamo DB instance</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/358/1*S9mD6rqSBroBVl-XCtGLqw.png" /><figcaption>Item Creation (Option1)</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/610/1*k6HDKoTntHYrycxCkPlgRg.png" /><figcaption>Item Creation (Option 2)</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/822/1*mEKBTGQrjSw5RF25YtEmsw.png" /><figcaption>Querying on the DB</figcaption></figure><h3>4. Data Integration Service</h3><h4><strong>i) AWS Glue</strong></h4><p>It is server-less data integration service that makes it easier to discover, prepare, move, and integrate data from multiple sources for analytics, machine learning, and application development</p><p>Used for ETL (Extract Transform &amp; Load)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/710/1*91Cmf8V6a415NwdQfNy4OQ.png" /><figcaption>Drag &amp; Drop Editor available in AWS Glue Studio</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/794/1*ngj3mLoKCMo7SF6o7F7QgA.png" /><figcaption>Py script editor available in AWS Glue Studio</figcaption></figure><p>AWS Glue Studio makes it easier to visually create, run, and monitor AWS Glue ETL jobs. You can build ETL jobs that move and transform data using a drag-and-drop editor, and AWS Glue automatically generates the code or we can even write custom py-Spark based scripts<br>We can start with a source, apply mapping transformations &amp; target in the editor shown on top which is the drag &amp; drop editor,</p><p>While on the bottom we see a py-Spark based script being used. We can write or upload our own spark code, or python shell script or use a Jupyter notebook for interactive development</p><h4>i) AWS Athena</h4><p>Amazon Athena is an interactive query service that makes it easy to analyze data in Amazon S3 using standard SQL. Athena is serverless, so there is no infrastructure to manage, and you pay only for the queries that you run.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bKOur3YI25L-JDXs9Dq17g.png" /><figcaption>AWS Athena</figcaption></figure><p>We simply point to our data in Amazon S3, define the schema, and start querying using standard SQL</p><p>We can also query on data lakes, warehouse &amp; other cloud &amp; on-premise sources</p><p>Athena is out-of-the-box integrated with <a href="https://aws.amazon.com/glue/">AWS Glue</a> Data Catalog, allowing us to create a unified metadata repository across various services</p><h3><strong>5. Serverless Architecture — AWS Lambda</strong></h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*egdyXNTe4cBbZPO2X6hrgg.png" /><figcaption>Overview of serverless architecture</figcaption></figure><p>Serverless architecture is an approach to software design that allows developers to build and run services without having to manage the underlying infrastructure.</p><p>Servers allow users to communicate with an application and access its business logic, but managing servers takes considerable time and resources. Teams have to maintain the server hardware, take care of software and security updates, and create backups in case of failure.</p><p>By adopting serverless architecture, developers can offload these responsibilities to a third-party provider, enabling them to focus on writing application code.</p><p>One of the most popular serverless architectures is Function as a Service (FaaS), where developers write their application code as a set of discrete functions. Each function will perform a specific task when triggered by an event, such as an incoming email or an HTTP request.</p><h4><strong>AWS Lambda</strong></h4><p>AWS Lambda is a serverless event driven service. You can run code without provisioning or managing infrastructure. Simply write and upload code as a .zip file or container image.</p><p><strong>Why use AWS Lambda? There are 3 advantages.</strong></p><ul><li>Lambda functions scale automatically, from a dozen events per day to hundreds of thousands per second</li><li>Save costs by paying only for the compute time you use — by per-millisecond — instead of provisioning infrastructure upfront for peak capacity</li><li>It will integrate with over 200 services and SaaS apps and you can access file system and databases</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*vifmdtmCJKvpP3J1aXEtSA.png" /><figcaption>Code snippet for Lambda function</figcaption></figure><p><strong>Example use-case :</strong></p><p>To give you an example of how a lambda can be used, let’s take a use case where when the user uploads a photo, it must be resized to fit different devices. Here we can use S3 buckets to store the images, and when a new image is uploaded, an event is triggered, and a lambda can be executed to convert the images and store back the images back to S3.</p><p>The below image shows how easy it is to write define a lambda function. We can choose different runtimes as well like Node JS, python or Golang.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*3hwvpd2nSl9MzZbqIpboaQ.png" /><figcaption>Lambda use-case illustration</figcaption></figure><p><strong>Limitations of Lambda :</strong></p><ul><li>Loss of control — you lack control over the software stack that your code runs on. If a hardware fault, data center outage, or other issue impacts one of your servers, you’re dependent on a cloud provider to fix it.</li><li>Cold starts are common in serverless environments, adding several seconds of latency to code execution when functions are invoked after a period of inactivity.</li><li>Not suitable for long running processes.</li><li>Max timeout — 15 minutes</li><li>Memory — 128 MB to 10,240 MB</li><li>Invocation payload — 6 MB (Synchronous) and 256 KB (Asynchronous)</li></ul><h3>6. Global Content Delivery Network — CloudFront</h3><p>Amazon CloudFront is a content delivery network operated by Amazon Web Services.</p><p>Content delivery networks provide a globally-distributed network of proxy servers that cache content, such as web videos or other bulky media, more locally to consumers, thus improving access speed for downloading the content.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/702/1*o_tIVkfOuC9XdlL06cyYdA.png" /><figcaption>CloudFront illustration</figcaption></figure><p>CloudFront improves read performance as content is cached at the edge</p><p>It provides DDoS protection (because worldwide), integrated with AWS Shield and AWS Web Application Firewall</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/718/1*rDGHMNMbwBX4Fl9jMZO8GA.png" /><figcaption>High Level view — internal working</figcaption></figure><h3>7. Cloud Monitoring — CloudWatch</h3><p>CloudWatch enables users to collect and view monitoring data for AWS infrastructures in a single platform</p><p>CloudWatch enables real-time monitoring of AWS resources such as Amazon Elastic Compute Cloud (<a href="https://www.techtarget.com/searchaws/definition/Amazon-Elastic-Compute-Cloud-Amazon-EC2">EC2</a>) instances, Amazon Elastic Block Store (EBS) volumes, Elastic Load Balancing and Amazon Relational Database Service (<a href="https://www.techtarget.com/searchaws/definition/Amazon-Relational-Database-Service-RDS">RDS</a>) instances.</p><p>The application automatically collects and provides metrics for CPU utilization, latency and request counts. Users can also stipulate additional metrics to be monitored, such as memory usage, transaction volumes or error rates.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/560/1*OlS_PhsB4VWpBQLJyN6NTg.png" /><figcaption>Billing Metric illustration</figcaption></figure><p>These were some of the broadly used AWS services and their brief introduction</p><p>Hope you found this article useful. Happy Learning!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c862c657807d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[VERTEX AI — (Part 1/2) Introduction and example use case]]></title>
            <link>https://medium.com/@anvitajoshi7/vertex-ai-part-1-2-introduction-and-example-use-case-45f0a50e152a?source=rss-4ac8149925f8------2</link>
            <guid isPermaLink="false">https://medium.com/p/45f0a50e152a</guid>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[vertex-ai]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[gcp]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Anvita Joshi]]></dc:creator>
            <pubDate>Wed, 02 Feb 2022 16:45:12 GMT</pubDate>
            <atom:updated>2022-02-16T18:41:25.683Z</atom:updated>
            <content:encoded><![CDATA[<h3>VERTEX AI — Introduction and example use case</h3><h4>What is Vertex AI?</h4><p>Vertex AI is Google Cloud’s end-to-end ML platform for data scientists and ML engineers to accelerate ML experimentation and deployment.</p><p>The platform unifies Google Cloud’s existing ML offerings into a single environment for efficiently building and managing the lifecycle of ML projects.</p><p>Google introduced Vertex AI in May 2021, it was announced in Google I/O (Innovation in the Open) held at Mountain View, California.</p><p>Features:</p><p>•Supports all open source frameworks — It supports all open source frameworks such as TensorFlow, PyTorch, and scikit-learn, along with custom containers for training and prediction.</p><p>•Unified UI for the entire ML workflow — It brings together the Google Cloud services for building ML under one unified UI and API. We can efficiently train and compare models using AutoML or custom code training &amp; A central model repository stores all our models that can deploy to the same endpoints</p><p>•Pre-trained APIs — Vertex AI has pre-trained APIs for vision, natural language, video and these could be incorporated into existing applications. We can also build new applications across use cases such as Translation and Speech to Text &amp; it has AutoML which allows developers to train high-quality models as per their business needs with a central registry for all datasets</p><ul><li>Integration — Developers can leverage BigQuery ML to create and execute machine learning models using standard SQL queries on existing tools and spreadsheets. Alternatively, we can export datasets from BigQuery into Vertex AI for integration across the data-to-AI life cycle.</li></ul><p>Benefits:</p><p>•Flexible</p><p>•Faster &amp; Auto-tuning hyperparameters (integrated with Vizer — Google’s AI Optimizer)</p><ul><li>Code reduction</li></ul><h4>Let’s jump to the cloud console and take a look at this</h4><p>Vertex AI <a href="https://console.cloud.google.com/vertex-ai">Dashboard</a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6k9reM93OcMsnoRz6qjSYQ.png" /></figure><p>Create <a href="https://console.cloud.google.com/vertex-ai/datasets">Dataset</a></p><p>We could create a variety of datasets depending on our use case :</p><ul><li>Image dataset for Image Classification (single or multi-label), Object detection and Image Segmentation</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uQsL1pkefAWdbWx1hVpUag.png" /></figure><ul><li>Tabular dataset for Regression or Classification and Forecasting</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rI30PYnyx4ZYV6d6BVxIfw.png" /></figure><ul><li>Text dataset for Text Classification (single or multi-label), Entity Extraction, Sentiment Analysis</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4znQHnu2LP-xw8_p3tMtMA.png" /></figure><ul><li>Video dataset for Action Recognition, Classification, Object Tracking</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tTol_CIk6k8PSAwNLwjTyQ.png" /></figure><h3>Example — Image Classification model using Vertex AI</h3><p>Dataset used — <a href="https://www.kaggle.com/alxmamaev/flowers-recognition">Flowers dataset</a> from Kaggle</p><p>Download the above dataset and store it in your local</p><p>Head to the cloud <a href="https://console.cloud.google.com/vertex-ai/datasets">console</a> into the Datasets section on the left and create the dataset</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zpXp8eM_oXDqBsNHgczXhg.png" /></figure><p>Select the dataset you created, we will now populate the training images into this</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*KXRF5EpYrhicwedLLuCB_Q.png" /></figure><p>Import each category of images seperately and label them accordingly</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*C7k_E_Trc2xT8qdqX7LxIw.png" /></figure><p>After labeling each image, Click on ‘TRAIN NEW MODEL’ option on the right</p><p>We have various training methods,</p><ul><li><em>AutoML</em> — Vertex AI’s inbuilt method which takes care of parameter tuning and choosing the best fit model depending on the type of data and the use case</li><li><em>Edge</em> is to deploy on-premises</li><li><em>Custom</em> training if we have our own code</li></ul><p>For now, we choose AutoML</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*keWHNHVzc-Psx0cnemabcA.png" /></figure><p>Model details — AutoML randomly splits the data into 80–10–10 for train-validate-test, we can change this based on our needs</p><p>Set 8 node-hours in Compute and Pricing section and start with the training</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*CzsR_HHpKU9TM3Yb2AfsFg.png" /></figure><p>You will receive an email notification once the training is complete,</p><p>You can now select your trained model and evaluate its metrics</p><p>We can see our model’s precision, the confusion matrix helps us understand where our model is confusing to results</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*prR67QM--wAOfwFFGsHOGA.png" /></figure><p>We can now deploy the model to an endpoint</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*eagjlsszegQK9paXeauYIA.png" /></figure><p>Set the number of compute nodes to 2 and Deploy</p><p>And finally once the endpoint gets created, we can test it by uploading an image</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tGL9GUMeZ136C3qN84JWow.png" /></figure><p>Alternatively, we can also make predictions using the SDK via REST API or Python</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*y194lA_iheoF8FRE-hEnlQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Loc7zXseYadvWc42_8RF5Q.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*14b3B0jsdKW5d1lPIHLx4w.png" /></figure><p>Congratulations, you’ve successfully created and deployed an Image Classification model using Vertex AI</p><p>Hope you found this article useful. Happy Learning!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=45f0a50e152a" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>