how to pass test data to my deployed model in azure ml through function app

Question

question

ranjithkumar-3800 asked Mar 4, '22 MayankBargali-MSFT edited Mar 28, '22

how to pass test data to my deployed model in azure ml through function app

i am trying to make predictions for the model which is deployed in azure ml through function app

1) my data is in azure blob storage
2) i want to pass that data to model ( which is deployed ) for predictions
3) once predictions is over i want to save that file again into blob storage

all above has to be done through function app

i dont have any idea on how to do it , any suggestions

azure-machine-learning azure-machine-learning-inference

ask1.png (72.6 KiB)

5 |1600 characters needed characters left characters exceeded

Attachments: Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

Answer 1 · 2022-03-04T12:47:51.377Z

ramr-msft answered Mar 4, '22

@ranjithkumar-3800 Thanks for the question. The deployment of a ML model to Azure Functions is in preview. Azure ML Managed Endpoints for enabling customers deploy models at cloud scale without needing to create & manage their own clusters. Here is link to the document that could help to Deploy and score a machine learning model by using an online endpoint (preview).
https://docs.microsoft.com/en-us/azure/machine-learning/how-to-deploy-managed-online-endpoints

· 1

5 |1600 characters needed characters left characters exceeded

Attachments: Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

ranjithkumar-3800 · Mar 04 at 12:59 PM

thanks for the rply @ramr-msft

for testing the mode which is deployed , using request library.. i followed this doc

https://docs.microsoft.com/en-us/azure/machine-learning/how-to-consume-web-service?tabs=python#call-the-service-python

which is working fine , can i do the same thing through function app using python , if yes may i know how to do that

thanks once again

0 ·

question