Deploy ML fashions in-built Amazon SageMaker Canvas to Amazon SageMaker real-time endpoints

Amazon SageMaker Canvas now helps deploying machine studying (ML) fashions to real-time inferencing endpoints, permitting you are taking your ML fashions to manufacturing and drive motion based mostly on ML-powered insights. SageMaker Canvas is a no-code workspace that permits analysts and citizen knowledge scientists to generate correct ML predictions for his or her enterprise wants.

Till now, SageMaker Canvas offered the flexibility to judge an ML mannequin, generate bulk predictions, and run what-if analyses inside its interactive workspace. However now you too can deploy the fashions to Amazon SageMaker endpoints for real-time inferencing, making it easy to eat mannequin predictions and drive actions exterior the SageMaker Canvas workspace. Being able to instantly deploy ML fashions from SageMaker Canvas eliminates the necessity to manually export, configure, take a look at, and deploy ML fashions into manufacturing, thereby saving lowering complexity and saving time. It additionally makes operationalizing ML fashions extra accessible to people, with out the necessity to write code.

On this publish, we stroll you thru the method to deploy a model in SageMaker Canvas to a real-time endpoint.

Overview of answer

For our use case, we’re assuming the function of a enterprise person within the advertising and marketing division of a cell phone operator, and we’ve efficiently created an ML mannequin in SageMaker Canvas to establish prospects with the potential danger of churn. Because of the predictions generated by our mannequin, we now wish to transfer this from our growth surroundings to manufacturing. To streamline the method of deploying our mannequin endpoint for inference, we instantly deploy ML fashions from SageMaker Canvas, thereby eliminating the necessity to manually export, configure, take a look at, and deploy ML fashions into manufacturing. This helps scale back complexity, saves time, and likewise makes operationalizing ML fashions extra accessible to people, with out the necessity to write code.

The workflow steps are as follows:

  1. Add a brand new dataset with the present buyer inhabitants into SageMaker Canvas. For the complete listing of supported knowledge sources, discuss with Import data into Canvas.
  2. Construct ML fashions and analyze their efficiency metrics. For directions, discuss with Build a custom model and Evaluate Your Model’s Performance in Amazon SageMaker Canvas.
  3. Deploy the approved model version as an endpoint for real-time inferencing.

You may carry out these steps in SageMaker Canvas with out writing a single line of code.


For this walkthrough, guarantee that the next stipulations are met:

  1. To deploy mannequin variations to SageMaker endpoints, the SageMaker Canvas admin should give the required permissions to the SageMaker Canvas person, which you’ll be able to handle within the SageMaker area that hosts your SageMaker Canvas software. For extra data, discuss with Permissions Management in Canvas.
  2. Implement the stipulations talked about in Predict customer churn with no-code machine learning using Amazon SageMaker Canvas.

You need to now have three mannequin variations educated on historic churn prediction knowledge in Canvas:

  • V1 educated with all 21 options and fast construct configuration with a mannequin rating of 96.903%
  • V2 educated with all 19 options (eliminated telephone and state options) and fast construct configuration and improved accuracy of 97.403%
  • V3 educated with customary construct configuration with 97.103% mannequin rating

Use the shopper churn prediction mannequin

Allow Present superior metrics on the mannequin particulars web page and overview the target metrics related to every mannequin model so as to choose the best-performing mannequin for deploying to SageMaker as an endpoint.

Based mostly on the efficiency metrics, we choose model 2 to be deployed.

Configure the mannequin deployment settings—deployment identify, occasion sort, and occasion depend.

As a place to begin, Canvas will robotically suggest the very best occasion sort and the variety of situations in your mannequin deployment. You may change it as per your workload wants.

You may take a look at the deployed SageMaker inference endpoint instantly from inside SageMaker Canvas.

You may change enter values utilizing the SageMaker Canvas person interface to deduce extra churn prediction.

Now let’s navigate to Amazon SageMaker Studio and take a look at the deployed endpoint.

Open a pocket book in SageMaker Studio and run the next code to deduce the deployed mannequin endpoint. Exchange the mannequin endpoint identify with your personal mannequin endpoint identify.

import boto3, sys
import pandas as pd

endpoint_name = "canvas-customer-churn-prediction-model"
sm_rt = boto3.Session().shopper('runtime.sagemaker')

payload = [['PA',163,806,403-2562, 'no', 'yes', 300, 8.16, 3, 7.57,3.93,4,6.5,4.07,100,5.11,4.92,6,5.67,3]]
physique = pd.DataFrame(payload).to_csv(header=False, index=False).encode("utf-8")

response = sm_rt.invoke_endpoint(EndpointName=endpoint_name, Physique=physique, ContentType="textual content/csv",Settle for="software/json")

response = response['Body'].learn().decode("utf-8")

Our unique mannequin endpoint is utilizing an ml.m5.xlarge occasion and 1 occasion depend. Now, let’s assume you count on the variety of end-users inferencing your mannequin endpoint will enhance and also you wish to provision extra compute capability. You may accomplish this instantly from inside SageMaker Canvas by selecting Replace configuration.

Clear up

To keep away from incurring future costs, delete the assets you created whereas following this publish. This contains logging out of SageMaker Canvas and deleting the deployed SageMaker endpoint. SageMaker Canvas payments you at some point of the session, and we suggest logging out of SageMaker Canvas whenever you’re not utilizing it. Check with Logging out of Amazon SageMaker Canvas for extra particulars.


On this publish, we mentioned how SageMaker Canvas can deploy ML fashions to real-time inferencing endpoints, permitting you are taking your ML fashions to manufacturing and drive motion based mostly on ML-powered insights. In our instance, we confirmed how an analyst can shortly construct a extremely correct predictive ML mannequin with out writing any code, deploy it on SageMaker as an endpoint, and take a look at the mannequin endpoint from SageMaker Canvas, in addition to from a SageMaker Studio pocket book.

To start out your low-code/no-code ML journey, discuss with Amazon SageMaker Canvas.

Particular due to everybody who contributed to the launch: Prashanth Kurumaddali, Abishek Kumar, Allen Liu, Sean Lester, Richa Sundrani, and Alicia Qi.

Concerning the Authors

Janisha Anand is a Senior Product Supervisor within the Amazon SageMaker Low/No Code ML workforce, which incorporates SageMaker Canvas and SageMaker Autopilot. She enjoys espresso, staying lively, and spending time together with her household.

Indy Sawhney is a Senior Buyer Options Chief with Amazon Net Companies. At all times working backward from buyer issues, Indy advises AWS enterprise buyer executives by their distinctive cloud transformation journey. He has over 25 years of expertise serving to enterprise organizations undertake rising applied sciences and enterprise options. Indy is an space of depth specialist with AWS’s Technical Subject Group for AI/ML, with specialization in generative AI and low-code/no-code Amazon SageMaker options.

Leave a Reply

Your email address will not be published. Required fields are marked *