none
Deployment issues with Azure ML

    Question

  • Hello I am facing issues when I am deploying my my Azure ML Experiments. in the predictive experiment its taking the unfiltered data  as input by default. its not  considering the select columns in data set operation. Does anyone know how to fix that problem?

    Thursday, January 31, 2019 9:24 PM

All replies

  • Hi, 

    Perhaps when you created the web service, it automatically set the web service input to one of the first modules, could you move your web service input to the module you? Then you need to re-run your experiment and publish the web service again. 

    Regards,
    Jaya


    Thursday, January 31, 2019 9:29 PM
  • thank you Jaya, I was using it for the first time so i didnt know if i can change it, Thanks! but now I have another issue, I am using Neural Network regression and my scored model is normalized. So now when I deploy it and want o test with with other data they  are non normalized, is there is a way around it? 
    Thursday, January 31, 2019 9:58 PM
  • Hi, 

    Since the model expects normalized data, you would need to do the preprocessing and normalize the data before scoring. 

    Regards,
    Jaya

    Friday, February 1, 2019 1:05 AM