CLI v2 managed online endpoint deployment YAML file referencing an existing inbuilt environment - how?

Question

The following documentation specifies this option...

To reference an existing environment, use the azureml:<environment-name>:<environment-version> syntax.

https://learn.microsoft.com/en-gb/azure/machine-learning/reference-yaml-deployment-managed-online#:~:text=To%20reference%20an%20existing%20environment%2C%20use%20the%20azureml%3A%3Cenvironment%2Dname%3E%3A%3Cenvironment%2Dversion%3E%20syntax.

Can anyone help me work out what I need to put into the YAML deployment config file please - tried multiple things and nothing works. I tried to highlight this over at the documentation GitHub but they told me to come here - not a documentation issue apparently! There are no example around.

This is my deployment YAML file - I want to reference the Azure Machine Learning existing environment which is called AzureML-sklearn-1.0-ubuntu20.04-py38-cpu version 32

$schema: https://azuremlschemas.azureedge.net/latest/managedOnlineDeployment.schema.json  
model: azureml:my-model:1  
code_configuration:  
  code: models/my-model/  
  scoring_script: score.py  
environment:   
 ???????  
  conda_file: models/my-model/conda.yml  
instance_type: Standard_DS2_v2  
instance_count: 2  
app_insights_enabled: true  

Thank you in advance

Answer 1

@Neil McAlister Thanks for the question. Here is the sample to create a Deployment YAML Definition.

$schema: https://azuremlschemas.azureedge.net/latest/managedOnlineDeployment.schema.json ​  
  
name: blue ​  
  
endpoint_name: my-endpoint ​  
  
model: ​  
  
    path: ../../model-1/model/ ​  
  
code_configuration: ​  
  
    code: ../../model-1/onlinescoring/ ​  
  
    scoring_script: score.py ​  
  
environment: ​  
  
    conda_file: ../../model-1/environment/conda.yml ​  
  
    image: mcr.microsoft.com/azureml/openmpi3.1.2-ubuntu18.04:20210727.v1 ​  
  
instance_type: Standard_DS2_v2 ​  
  
instance_count: 1  

Answer 2

Hello @Ramr-msft and thank you for your reply

I found that using your YAML it will create a repository for it in the connected Azure Container Registry

When using the GUI to do the same operation though, it does not - this is preferred to keep costs under control

I have found using the following environment statement works and a repository is not created in the Container Register, however I am unable to provide the conda yaml file in this definition as it fails syntax checks.

 $schema: https://azuremlschemas.azureedge.net/latest/managedOnlineDeployment.schema.json  
 model: azureml:my-model:1  
 code_configuration:  
   code: models/my-model/  
   scoring_script: score.py  
environment: azureml:AzureML-sklearn-1.0-ubuntu20.04-py38-cpu:32  
instance_type: Standard_DS2_v2  
instance_count: 2  
app_insights_enabled: true  

Give it a try and see the difference between GUI and YAML deployments using an existing environment with a conda requirements file

Thanks

Neil