cancel
Showing results for 
Search instead for 
Did you mean: 

Deploy a DataRobot model with API

Deploy a DataRobot model with API

Hello, 

 

I would like to deploy a DataRobot learning model using the API (Python).

 

In the documentation, I saw the function create_from_learning_model

This function has the following parameters : 

  • model_id
  • label
  • description
  • default_prediction_server_i
  • importance
  • prediction_threshold
  • status

But there is no parameters for :

  • Set association ID and enable target monitoring
  • Enable feature drift tracking
  • Enable prediction rows storage for challenger analysis
  • Enable automatic actuals feedback for time series models
  • Track attributes for segmented analysis of training data and predictions
  • Settings monitoring (drift metric, drift threshold, accuracy metric, accuracy threshold ...)

Furthermore, there is no functions to submit retraining data.

 

In a mature project, it's essential to be able to automate all this without using the UI. For instance, if I want to retrain a model every week, I will have to upload by hand the new training data set using the UI ? 

 

Will the access to all these settings be available soon?

 

Another question : is it possible to duplicate an existing deployment ?

 

Thanks !

 

 

 

 

Labels (2)
0 Kudos
1 Solution
2 Replies