Managed online endpoints SKU list
The following table shows the virtual machine (VM) stock keeping units (SKUs) that are supported for Azure Machine Learning managed online endpoints. Each SKU is a unique alphanumeric code assigned to a particular VM that can be purchased.
The full SKU names listed in the table can be used for Azure CLI or Azure Resource Manager templates (ARM templates) requests to create and update deployments.
For more information on configuration details such as CPU and RAM, see Azure Machine Learning Pricing and VM sizes.
Relative Size | General Purpose | Compute Optimized | Memory Optimized | GPU |
---|---|---|---|---|
X-Small | Standard_DS1_v2 Standard_DS2_v2 Standard_D2a_v4 Standard_D2as_v4 |
Standard_F2s_v2 | Standard_E2s_v3 | Standard_NC4as_T4_v3 |
Small | Standard_DS3_v2 Standard_D4a_v4 Standard_D4as_v4 |
Standard_F4s_v2 Standard_FX4mds |
Standard_E4s_v3 | Standard_NC6s_v2 Standard_NC6s_v3 Standard_NC8as_T4_v3 |
Medium | Standard_DS4_v2 Standard_D8a_v4 Standard_D8as_v4 |
Standard_F8s_v2 Standard_FX12mds |
Standard_E8s_v3 | Standard_NC12s_v2 Standard_NC12s_v3 Standard_NC16as_T4_v3 |
Large | Standard_DS5_v2 Standard_D16a_v4 Standard_D16as_v4 |
Standard_F16s_v2 | Standard_E16s_v3 | Standard_NC24s_v2 Standard_NC24s_v3 Standard_NC64as_T4_v3 Standard_NC24ads_A100_v4 |
X-Large | Standard_D32a_v4 Standard_D32as_v4 Standard_D48a_v4 Standard_D48as_v4 Standard_D64a_v4 Standard_D64as_v4 Standard_D96a_v4 Standard_D96as_v4 |
Standard_F32s_v2 Standard_F48s_v2 Standard_F64s_v2 Standard_F72s_v2 Standard_FX24mds Standard_FX36mds Standard_FX48mds |
Standard_E32s_v3 Standard_E48s_v3 Standard_E64s_v3 |
Standard_NC48ads_A100_v4 Standard_NC96ads_A100_v4 Standard_ND96asr_v4 Standard_ND96amsr_A100_v4 Standard_ND40rs_v2 |
Caution
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 or ReourceNotReady errors, 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.
Note
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. VM SKUs that are exempted from this extra quota reservation are listed below:
- Standard_NC24ads_A100_v4
- Standard_NC48ads_A100_v4
- Standard_NC96ads_A100_v4
- Standard_ND96asr_v4
- Standard_ND96amsr_A100_v4
- Standard_ND40rs_v2