AWS general documentation change
Summary
Added ml.p5.4xlarge instance quotas for reserved capacities and training plans
Security assessment
This establishes compute resource reservation limits for specific instance types, focusing on workload availability rather than security features or vulnerabilities.
Diff
diff --git a/general/latest/gr/sagemaker.md b/general/latest/gr/sagemaker.md index 7d032a62d..65abd211d 100644 --- a//general/latest/gr/sagemaker.md +++ b//general/latest/gr/sagemaker.md @@ -176,0 +177 @@ Maximum subsampled dataset size AutoML job can be run on | Each supported Region +Maximum total FSx Lustre capacity in TiB across SageMaker HyperPod restricted instance groups | Each supported Region: 0 | [Yes](https://console.aws.amazon.com/servicequotas/home/services/sagemaker/quotas/L-FC4D7F52) | Maximum total FSx Lustre capacity in TiB across SageMaker HyperPod restricted instance groups @@ -187,0 +189 @@ Number of ml.p5.48xlarge instances in reserved capacity across training plans pe +Number of ml.p5.4xlarge instances in reserved capacity across training plans per Region | Each supported Region: 0 | [Yes](https://console.aws.amazon.com/servicequotas/home/services/sagemaker/quotas/L-F737A0D0) | The maximum number of ml.p5.4xlarge instances available in SageMaker reserved capacities across all training plans in this account and Region at any given time. These reservations secure compute resources across multiple Availability Zones, ensuring availability for large-scale AI model training workloads. @@ -1549,0 +1552,3 @@ ml.p5.48xlarge for training warm pool usage | Each supported Region: 0 | [Yes]( +ml.p5.4xlarge for processing job usage | Each supported Region: 0 | [Yes](https://console.aws.amazon.com/servicequotas/home/services/sagemaker/quotas/L-A60122FA) | ml.p5.4xlarge for processing job usage +ml.p5.4xlarge for training job usage | Each supported Region: 0 | [Yes](https://console.aws.amazon.com/servicequotas/home/services/sagemaker/quotas/L-C2764A80) | ml.p5.4xlarge for training job usage +ml.p5.4xlarge for training warm pool usage | Each supported Region: 0 | [Yes](https://console.aws.amazon.com/servicequotas/home/services/sagemaker/quotas/L-E511AAB3) | ml.p5.4xlarge for training warm pool usage