Training Type of Personalize solution version in the console is Manual even when automatic training was set for the solution

Greeting

Hi Mustapha!

Thank you for bringing up this issue with Amazon Personalize! It sounds like you’ve run into something unexpected, and I’m happy to help clarify this for you. 😊

---

Clarifying the Issue

You mentioned that the training type of your solution version appears as "manual" in the console, even though you configured it for automatic training during the creation process. It’s understandable to be concerned about whether automatic retraining will occur as intended. Let’s work together to determine if this is a console display issue, a configuration oversight, or a potential bug. I’ll guide you through the steps to investigate and resolve it.

---

Key Terms

• Training Configuration: Defines how and when a model is retrained in Amazon Personalize.
• Automatic Training: A feature that ensures models are retrained periodically based on new data.
• Solution Version: A trained model version created as part of an Amazon Personalize solution.

---

The Solution (Our Recipe)

Steps at a Glance:

1. Verify the training configuration settings in the console or programmatically.
2. Check for recent retraining activity in the solution version details.
3. Test automatic training with a new solution version to confirm behavior.
4. Contact AWS Support if the issue persists.

---

Step-by-Step Guide:

1. Verify the training configuration settings in the console or programmatically:

   • In the AWS Management Console, navigate to the solution details and check if the training configuration is set to "automatic."

   • Alternatively, use this Python script with Boto3 to retrieve the training configuration programmatically:

     import boto3
     
     personalize = boto3.client('personalize')
     
     try:
         response = personalize.describe_solution(
             solutionArn='arn:aws:personalize:region:account-id:solution/solution-name'
         )
         training_mode = response['solution']['trainingMode']
         print(f"Training mode: {training_mode}")
     except Exception as e:
         print(f"Error retrieving solution details: {e}")

     Make sure you have the AWS SDK for Python (boto3) installed and configured with appropriate credentials.

---

2. Check for recent retraining activity in the solution version details:

   • Use this script to list solution versions and their statuses:

     try:
         response = personalize.list_solution_versions(
             solutionArn='arn:aws:personalize:region:account-id:solution/solution-name'
         )
         for version in response['solutionVersions']:
             print(f"Version ARN: {version['solutionVersionArn']}, Status: {version['status']}")
     except Exception as e:
         print(f"Error listing solution versions: {e}")

   • Look for timestamps or logs indicating when the model was last retrained. This can help verify if retraining occurred despite the display issue.

---

3. Test automatic training with a new solution version to confirm behavior:
   • Create a new solution with automatic training explicitly enabled:

     try:
         response = personalize.create_solution(
             name='test-solution',
             datasetGroupArn='arn:aws:personalize:region:account-id:dataset-group/dataset-group-name',
             performAutoML=True,
             performHPO=False
         )
         print(f"New Solution ARN: {response['solutionArn']}")
     except Exception as e:
         print(f"Error creating solution: {e}")

---

4. Contact AWS Support if the issue persists:
   • If you continue to see inconsistencies, contact AWS Support. Provide them with screenshots, solution ARNs, and any error messages from your testing.

---

Closing Thoughts

Issues like this can sometimes stem from display discrepancies in the AWS Management Console. However, by verifying your training configurations both visually and programmatically, you can determine whether this is simply a UI issue or something deeper. Let me know if you’d like more guidance on debugging or troubleshooting this further! 🚀

---

Farewell

Take care, and good luck with your Amazon Personalize projects, Mustapha! 😊

---

Cheers,

Aaron 😊