I am trying to run some code to train a model, while logging my results to MLflow on Databricks. I keep getting the following error when I try to make a call to mlflow.set_experiment()
,
raise ValueError('Enum {} has no value defined for name {!r}'.format(
ValueError: Enum ErrorCode has no value defined for name '403'
What exactly is going on here?
I am using Databricks Connect to run my code and the section where the error pops up looks like this,
# set remote tracking server URI
mlflow.set_tracking_uri(remote_server_uri)
# create the MLflow client
client = MlflowClient(remote_server_uri)
# set experiment to log mlflow runs
mlflow.set_experiment(experiment_name)
It looks like the reason I was not able to log my experiments to MLflow is because I have missed the configuration steps that need to be performed before this can be achieved.
The document given below is a good guide on what needs to be done,
https://learn.microsoft.com/en-us/azure/databricks/applications/mlflow/access-hosted-tracking-server
In essence, the following steps must be performed when trying to access the MLflow tracking server from outside of Databricks,