Skip to content

Commit 5a0d2d5

Browse files
authored
Merge pull request #378 from geffner/patch-1
Update reference-managed-online-endpoints-vm-sku-list.md
2 parents 27a694d + 5ab424e commit 5a0d2d5

File tree

1 file changed

+1
-1
lines changed

1 file changed

+1
-1
lines changed

articles/machine-learning/reference-managed-online-endpoints-vm-sku-list.md

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -104,7 +104,7 @@ The following table shows the virtual machine (VM) stock keeping units (SKUs) th
104104
| standardNDv5H100Family | STANDARD_ND96ISR_H100_v5 | Yes | NvidiaGpu | 8 | 96 | Yes |
105105

106106
> [!CAUTION]
107-
> Small VM SKUs such as `Standard_DS1_v2` and `Standard_F2s_v2` may be too small for bigger models and may lead to container termination due to insufficient memory, not enough space on the disk, or probe failure as it takes too long to initiate the container. If you face [OutOfQuota errors](how-to-troubleshoot-online-endpoints.md?tabs=cli#error-outofquota) or [ReourceNotReady errors](how-to-troubleshoot-online-endpoints.md?tabs=cli#error-resourcenotready), try bigger VM SKUs. If you want to reduce the cost of deploying multiple models with managed online endpoint, see [Deployment for several local models](concept-online-deployment-model-specification.md#deployment-for-several-local-models).
107+
> Small VM SKUs such as `Standard_DS1_v2` and `Standard_F2s_v2` may be too small for bigger models and may lead to container termination due to insufficient memory, not enough space on the disk, or probe failure as it takes too long to initiate the container. If you face [OutOfQuota errors](how-to-troubleshoot-online-endpoints.md?tabs=cli#error-outofquota) or [ResourceNotReady errors](how-to-troubleshoot-online-endpoints.md?tabs=cli#error-resourcenotready), try bigger VM SKUs. If you want to reduce the cost of deploying multiple models with managed online endpoint, see [Deployment for several local models](concept-online-deployment-model-specification.md#deployment-for-several-local-models).
108108
109109
> [!NOTE]
110110
> We recommend having more than 3 instances for deployments in production scenarios. In addition, Azure Machine Learning reserves 20% of your compute resources for performing upgrades on some VM SKUs as described in [Virtual machine quota allocation for deployment](how-to-manage-quotas.md#virtual-machine-quota-allocation-for-deployment). VM SKUs that are exempted from this extra quota reservation are specified in the "Skip 20% Reservation" column.

0 commit comments

Comments
 (0)