SKU-lista över hanterade slutpunkter online
I följande tabell visas de lagerhållningsenheter för virtuella datorer (VM) som stöds för Azure Machine Learning-hanterade onlineslutpunkter. Varje SKU är en unik alfanumerisk kod som tilldelats till en viss virtuell dator som kan köpas.
De fullständiga SKU-namnen som anges i tabellen kan användas för Azure CLI- eller Azure Resource Manager-mallar (ARM-mallar) för att skapa och uppdatera distributioner.
Mer information om konfigurationsinformation som CPU och RAM finns i Prissättning för Azure Machine Learning och VM-storlekar.
Efternamn | Namn på VM-storlek | Stöder Infiniband | Arkitektur | numberOfGPUs | numberOfCores | Hoppa över 20 % reservation |
---|---|---|---|---|---|---|
standardDASv4Family | STANDARD_D2AS_V4 | - | Processor | 0 | 2 | - |
standardDASv4Family | STANDARD_D4AS_V4 | - | Processor | 0 | 4 | - |
standardDASv4Family | STANDARD_D8AS_V4 | - | Processor | 0 | 8 | - |
standardDASv4Family | STANDARD_D16AS_V4 | - | Processor | 0 | 16 | - |
standardDASv4Family | STANDARD_D32AS_V4 | - | Processor | 0 | 32 | - |
standardDASv4Family | STANDARD_D48AS_V4 | - | Processor | 0 | 48 | - |
standardDASv4Family | STANDARD_D64AS_V4 | - | Processor | 0 | 64 | - |
standardDASv4Family | STANDARD_D96AS_V4 | - | Processor | 0 | 96 | - |
standardDAv4Family | STANDARD_D2A_V4 | - | Processor | 0 | 2 | - |
standardDAv4Family | STANDARD_D4A_V4 | - | Processor | 0 | 4 | - |
standardDAv4Family | STANDARD_D8A_V4 | - | Processor | 0 | 8 | - |
standardDAv4Family | STANDARD_D16A_V4 | - | Processor | 0 | 16 | - |
standardDAv4Family | STANDARD_D32A_V4 | - | Processor | 0 | 32 | - |
standardDAv4Family | STANDARD_D48A_V4 | - | Processor | 0 | 48 | - |
standardDAv4Family | STANDARD_D64A_V4 | - | Processor | 0 | 64 | - |
standardDAv4Family | STANDARD_D96A_V4 | - | Processor | 0 | 96 | - |
standardDSv2Family | STANDARD_DS1_V2 | - | Processor | 0 | 1 | - |
standardDSv2Family | STANDARD_DS2_V2 | - | Processor | 0 | 2 | - |
standardDSv2Family | STANDARD_DS3_V2 | - | Processor | 0 | 4 | - |
standardDSv2Family | STANDARD_DS4_V2 | - | Processor | 0 | 8 | - |
standardDSv2Family | STANDARD_DS5_V2 | - | Processor | 0 | 16 | - |
standardESv3Family | STANDARD_E2S_V3 | - | Processor | 0 | 2 | - |
standardESv3Family | STANDARD_E4S_V3 | - | Processor | 0 | 4 | - |
standardESv3Family | STANDARD_E8S_V3 | - | Processor | 0 | 8 | - |
standardESv3Family | STANDARD_E16S_V3 | - | Processor | 0 | 16 | - |
standardESv3Family | STANDARD_E32S_V3 | - | Processor | 0 | 32 | - |
standardESv3Family | STANDARD_E48S_V3 | - | Processor | 0 | 48 | - |
standardESv3Family | STANDARD_E64S_V3 | - | Processor | 0 | 64 | - |
standardFSv2Family | STANDARD_F2S_V2 | - | Processor | 0 | 2 | - |
standardFSv2Family | STANDARD_F4S_V2 | - | Processor | 0 | 4 | - |
standardFSv2Family | STANDARD_F8S_V2 | - | Processor | 0 | 8 | - |
standardFSv2Family | STANDARD_F16S_V2 | - | Processor | 0 | 16 | - |
standardFSv2Family | STANDARD_F32S_V2 | - | Processor | 0 | 32 | - |
standardFSv2Family | STANDARD_F48S_V2 | - | Processor | 0 | 48 | - |
standardFSv2Family | STANDARD_F64S_V2 | - | Processor | 0 | 64 | - |
standardFSv2Family | STANDARD_F72S_V2 | - | Processor | 0 | 72 | - |
standardFXMDVSFamily | STANDARD_FX4MDS | - | Processor | 0 | 4 | - |
standardFXMDVSFamily | STANDARD_FX12MDS | - | Processor | 0 | 12 | - |
standardFXMDVSFamily | STANDARD_FX24MDS | - | Processor | 0 | 24 | - |
standardFXMDVSFamily | STANDARD_FX36MDS | - | Processor | 0 | 36 | - |
standardFXMDVSFamily | STANDARD_FX48MDS | - | Processor | 0 | 48 | - |
standardLASv3Family | STANDARD_L8AS_V3 | - | Processor | 0 | 8 | - |
standardLASv3Family | STANDARD_L16AS_V3 | - | Processor | 0 | 16 | - |
standardLASv3Family | STANDARD_L32AS_V3 | - | Processor | 0 | 32 | - |
standardLASv3Family | STANDARD_L48AS_V3 | - | Processor | 0 | 48 | - |
standardLASv3Family | STANDARD_L64AS_V3 | - | Processor | 0 | 64 | - |
standardLASv3Family | STANDARD_L80AS_V3 | - | Processor | 0 | 80 | - |
standardLSv2Family | STANDARD_L8S_V2 | - | Processor | 0 | 8 | - |
standardLSv2Family | STANDARD_L16S_V2 | - | Processor | 0 | 16 | - |
standardLSv2Family | STANDARD_L32S_V2 | - | Processor | 0 | 32 | - |
standardLSv2Family | STANDARD_L48S_V2 | - | Processor | 0 | 48 | - |
standardLSv2Family | STANDARD_L64S_V2 | - | Processor | 0 | 64 | - |
standardLSv2Family | STANDARD_L80S_V2 | - | Processor | 0 | 80 | - |
standardLSv3Family | STANDARD_L8S_V3 | - | Processor | 0 | 8 | - |
standardLSv3Family | STANDARD_L16S_V3 | - | Processor | 0 | 16 | - |
standardLSv3Family | STANDARD_L32S_V3 | - | Processor | 0 | 32 | - |
standardLSv3Family | STANDARD_L48S_V3 | - | Processor | 0 | 48 | - |
standardLSv3Family | STANDARD_L64S_V3 | - | Processor | 0 | 64 | - |
standardLSv3Family | STANDARD_L80S_V3 | - | Processor | 0 | 80 | - |
standardNCADSA100v4Family | STANDARD_NC24ADS_A100_V4 | - | NvidiaGpu | 1 | 24 | Ja |
standardNCADSA100v4Family | STANDARD_NC48ADS_A100_V4 | - | NvidiaGpu | 2 | 48 | Ja |
standardNCADSA100v4Family | STANDARD_NC96ADS_A100_V4 | - | NvidiaGpu | 4 | 96 | Ja |
Standard-NCASv3_T4-familj | STANDARD_NC4AS_T4_V3 | - | NvidiaGpu | 1 | 4 | - |
Standard-NCASv3_T4-familj | STANDARD_NC8AS_T4_V3 | - | NvidiaGpu | 1 | 8 | - |
Standard-NCASv3_T4-familj | STANDARD_NC16AS_T4_V3 | - | NvidiaGpu | 1 | 16 | - |
Standard-NCASv3_T4-familj | STANDARD_NC64AS_T4_V3 | - | NvidiaGpu | 4 | 64 | - |
standardNCSv2Family | STANDARD_NC6S_V2 | - | NvidiaGpu | 1 | 6 | - |
standardNCSv2Family | STANDARD_NC12S_V2 | - | NvidiaGpu | 2 | 12 | - |
standardNCSv2Family | STANDARD_NC24S_V2 | - | NvidiaGpu | 4 | 24 | - |
standardNCSv3Family | STANDARD_NC6S_V3 | - | NvidiaGpu | 1 | 6 | - |
standardNCSv3Family | STANDARD_NC12S_V3 | - | NvidiaGpu | 2 | 12 | - |
standardNCSv3Family | STANDARD_NC24S_V3 | - | NvidiaGpu | 4 | 24 | - |
standardNCADSH100v5Family | STANDARD_NC40ADS_H100_V5 | - | NvidiaGpu | 1 | 40 | Ja |
standardNCADSH100v5Family | STANDARD_NC80ADIS_H100_V5 | - | NvidiaGpu | 2 | 80 | Ja |
standard NDAMSv4_A100Family | STANDARD_ND96AMSR_A100_V4 | Ja | NvidiaGpu | 8 | 96 | Ja |
Standard-NDASv4_A100-familj | STANDARD_ND96ASR_V4 | Ja | NvidiaGpu | 8 | 96 | Ja |
standardNDSv2Family | STANDARD_ND40RS_V2 | Ja | NvidiaGpu | 8 | 40 | Ja |
standardNDv5H100Family | STANDARD_ND96IS_H100_v5 | - | NvidiaGpu | 8 | 96 | Ja |
standardNDv5H100Family | STANDARD_ND96ISR_H100_v5 | Ja | NvidiaGpu | 8 | 96 | Ja |
Varning
Små VM-SKU:er som Standard_DS1_v2
och Standard_F2s_v2
kan vara för små för större modeller och kan leda till att containern avslutas på grund av otillräckligt minne, inte tillräckligt med utrymme på disken eller avsökningsfel eftersom det tar för lång tid att initiera containern. Om du stöter på OutOfQuota-fel eller ReourceNotReady-fel kan du prova större VM-SKU:er. Om du vill minska kostnaden för att distribuera flera modeller med hanterad onlineslutpunkt kan du läsa Distribution för flera lokala modeller.
Kommentar
Vi rekommenderar att du har fler än 3 instanser för distributioner i produktionsscenarier. Dessutom reserverar Azure Machine Learning 20 % av dina beräkningsresurser för att utföra uppgraderingar på vissa VM-SKU:er enligt beskrivningen i kvotallokering för virtuella datorer för distribution. VM-SKU:er som är undantagna från den här extra kvotreservationen anges i kolumnen "Hoppa över 20 % reservation".