Datasaur Dinamic with Amazon SageMaker Autopilot

Introduction

Datasaur Dinamic with Amazon SageMaker lets you build and train models using Datasaur and AWS together. Your labeled data is directly connected to Amazon SageMaker to start the training process.

Quick start guide

Follow these steps to train a model with Datasaur Dinamic and Amazon SageMaker.

Setup Datasaur Dinamic extension

1. Create a custom row labeling project.

2. If the dataset is unlabeled, start by labeling part of the data first. For example, if the dataset has 20 rows, you can label 10–15 rows and automate the rest later.

3. Enable the Datasaur Dinamic extension from the Manage extensions dialog.

   
4. Select the AWS Sagemaker provider.

   

Setup an IAM role in AWS

Make sure you have an AWS S3 bucket available for storing training data and model outputs.

Create IAM Policies

1. Open the IAM page and navigate to the Policies section.

2. Click Create policy.

   
3. Create the following policies:

   • sagemaker-limited-access

   • sagemaker-s3-specific-permissions

Use the provided policy examples and update the variables with your bucket information.

sagemaker-limited-access.txt

3KB

Open

sagemaker-limited-access

sagemaker-s3-specific-permissions.txt

1KB

Open

sagemaker-s3-specific-permissions

Create an IAM role

1. Go to the IAM page and navigate to the Roles section.

2. Click on Create role.

3. Select AWS account as the role type and enter the Account ID and External ID generated in the Datasaur Dinamic extension.

   

   Select trusted entity
4. In the Add permissions step, attach the policies created earlier.

   

   Add permissions
5. In step 3, complete the role creation process by entering a role name and details.

6. After creating the role, open the role and go to the Trust relationships tab.

7. Edit the trust policy and add the following statement:

{
    "Effect": "Allow",
    "Principal": {
        "Service": "sagemaker.amazonaws.com"
    },
    "Action": "sts:AssumeRole"
}

Edit trust policy

1. After saving the changes, copy the Role ARN from the role summary and paste it into the Datasaur Dinamic extension.

   

   Role summary

Start training

1. In the Datasaur Dinamic extension, enter the AWS S3 Bucket Name and the Role ARN.

   

   Datasaur Dinamic with Amazon Sagemaker
2. Click Train to start training with Amazon SageMaker Autopilot.

   

   Datasaur Dinamic training progress
3. After training is complete, the model name will appear in the extension. You can also find the trained model in your Amazon SageMaker model list.

   

   Datasaur Dinamic model name