SageMaker Training Plans now allows seamless extension of capacity commitments

Amazon SageMaker has introduced a new feature for its Training Plans, allowing users to extend their existing GPU capacity commitments without needing to reconfigure their workloads. This update is particularly beneficial for AI workloads that require more time than initially planned, ensuring continuous access to the necessary capacity.

SageMaker Training Plans enable users to reserve GPU capacity for specific time frames with cluster sizes of up to 64 instances. With the new extension capability, users can prolong their plans in 1-day increments for up to 14 days, or in 7-day increments for as long as 182 days (26 weeks). These extensions can be conveniently initiated via the SageMaker console or through an API.

Once an extension is purchased, the workload continues to operate seamlessly without any need for manual reconfiguration. This feature enhances the flexibility and efficiency of managing AI workloads on SageMaker, allowing users to optimize their training plans according to their timelines and budgetary constraints.

SageMaker AI automatically provisions the necessary infrastructure and executes the AI workloads on the allocated compute resources, eliminating the need for manual intervention. Users can consult the SageMaker AI pricing page for detailed information about instance availability across different AWS Regions.

For further details on how to extend your training plans, refer to the Amazon SageMaker Training Plans User Guide.