Partial Dependency plot of a lab model
Mohammed
Dataiku DSS Core Designer, Dataiku DSS ML Practitioner, Registered Posts: 43 ✭✭✭
Hi,
I am working on building a prediction model in Dataiku. I am building models in the lab, and I'm trying to build a dashboard that can compare different versions of the model in LAB.
I'm accessing different versions of the model and comparing the different aspects of the model.
I'm accessing the Lab model as follows.
import dataiku
from dataiku.doctor.posttraining.model_information_handler import PredictionModelInformationHandler
fmi = "FULL_MODEL_ID"
model_handler = PredictionModelInformationHandler.from_full_model_id(fmi)
Is there any way to access the partial dependence of this LAB model via API similar to the saved model as given below ?
model_test = p.get_saved_model("SAVED_MODEL_ID")
details = model_test.get_version_details("VERSION_ID")
details.details.get('iperf', {})["partialDependencies"]
Operating system used: Windows
Tagged: