1

I try to use azuremlsdk to deploy a locally trained model (a perfectly valid use case AFIK). I follow this and managed to create a ML workspace and register a "model" like so:

library(azuremlsdk)

interactive_auth <- interactive_login_authentication(tenant_id="xxx")
ws <- get_workspace(
        name = "xxx", 
        subscription_id = "xxx", 
        resource_group ="xxx", 
        auth = interactive_auth
)

add <- function(a, b) {
    return(a + b)
}

add(1,2)

saveRDS(add, file = "D:/add.rds")

model <- register_model(
    ws, 
    model_path = "D:/add.rds", 
    model_name = "add_model",
    description = "An amazing model"
)

This seemed to work fine, as I get some nice log messages telling me that the model was registered. For my sanity, I wonder where can I find this registered ("materialised") model/object/function in the Azure UI please?

cs0815
  • 16,751
  • 45
  • 136
  • 299

1 Answers1

1

On ml.azure.com, there is a "Models" option on the left-hand blade.

UI Sidebar

Anders Swanson
  • 3,637
  • 1
  • 18
  • 43
  • oh thanks. I was looking at portal.azure.com – cs0815 May 14 '21 at 14:29
  • 1
    yeah the UI used to be in the portal, but it's been migrated to it's own wesbite, which is great. – Anders Swanson May 14 '21 at 14:41
  • 1
    Is good to have found the ui. at the moment I try to do everything via code and am stuck here: https://stackoverflow.com/questions/67535014/deploy-model-to-azure-machine-learning-via-azuremlsdk if you have any input to this one, that would be very much appreciated (-: I might try the ui way first - maybe it can insure code which i can use in a pipeline .... – cs0815 May 14 '21 at 14:43