Deploy Azure ML Designer Web Service model

Question

Im trying to run R code to detect outliers in datasets. Im trying to publish as a webservice so it can be used to pre process data.

In Azure ML Designer,I have a Execute R script which contains my model.

When I try to publish as a web service I get the error "Cannot create inference because there is no model on this pipeline". I see python has a Create Python module.

Can I use an Execute R Script to create a webservice with my model?

Im good at R and still learning Azure and would really appreciate some advice on. the above.

hi Vincent. Did you add the data source and other necessary components to create a pipeline in the Azure ML Designer? — NITHIN B
Hi Nithin. My pipeline runs perfectly , however I don't have the model registered before including in my pipeline. I think I have to go back register the model and add that to my pipeline. In the meantime Ive been looking at other ways, Ive been looking at running the script via azuremlsdk package and a compute instance of R from Azure ML. Any thoughts much appreciated. — VincentT

Lu Zhang Lu Zhang · Accepted Answer · 2020-04-24T13:05:59

Could you please follow the steps below to deploy the web service in Designer:

Manually drag "web service input" & "web service output" to the graph. 1st Note: Designer currently needs graph contains web service input & web service output both, then support realtimeEndpoint deploy.

2nd Note: If you found your graph contains "web service input" & "web service output", but no "Deploy" button, and only see below three buttons. => please follow below highlights to Clone this graph as a new pipeline, and retry.

Deploy Azure ML Designer Web Service model

1 Answers