Is there a library of SageMaker HyperPod Recipes I can choose from based on the need of my applications?

Hi,

Yes, the library of SageMaker recipes is available in the form of a Git repo that you can browse: https://github.com/aws/sagemaker-hyperpod-recipes

You will find all details in this blog post that is detailing the launch of SageMaker Recipes: https://aws.amazon.com/blogs/aws/accelerate-foundation-model-training-and-fine-tuning-with-new-amazon-sagemaker-hyperpod-recipes/

Best,

Didier

Yes, Amazon SageMaker HyperPod offers a library of predefined recipes that you can choose from based on your application needs. Predefined Recipes: SageMaker HyperPod provides a collection of pre-configured recipes for various popular foundation models, including Llama 3.1, Llama 3.2, Mistral, and Mixtral.You can create custom recipes tailored to your specific requirements. This allows you to define your own training configurations and add them to your internal library. Custom recipes can be saved and reused across your organization, promoting consistency and efficiency in your machine learning workflows.

The recipes automate critical steps such as loading training datasets, applying distributed training techniques, and managing the end-to-end training loop. This automation reduces manual work and potential errors.

See more in the documentation and the GitRepo.