Pycaret MlFlow authentication

Question

How can I use log_environment = True in Pycaret setup with

import os import mlflow mlflow.set_tracking_uri("https://dagshub.com/BexTuychiev/pet_pawpularity.mlflow") os.environ["MLFLOW_TRACKING_USERNAME"] = "MLFLOW_TRACKING_USERNAME" os.environ["MLFLOW_TRACKING_PASSWORD"] = "MLFLOW_TRACKING_PASSWORD"

Without getting RestException: INTERNAL_ERROR: Response: {'error': 'not found'}

Answer 1

In case others may run into this challege. When using DAGsHub, the tentative solution is creating a experiment in mlflow ui first.

In my situation the experiment name was called default. So, I had to make sure it ws reflected as experiment_name = 'Default' in the setup().

The wierd thing that I experienced was it would not create new experiment name using the pycaret setup() and I recieved the error mentioned in the above post.