Run with UUID *** is already active when running automl

ALIDI — Mon, 12 Dec 2022 15:02:10 GMT

Hi,

I'm tried using databricks autoML API following the documentation and example notebook. The documentation and example are pretty straight forward however I encountered the following error:

Exception: Run with UUID 1315376a0cbb4657b5d23fa552efba4b is already active. To start a new run, first end the current run with mlflow.end_run(). To start a nested run, call start_run with nested=True

The code I used:

mlflow.end_run()
mlflow.start_run(nested=True,run_name='autoML')
mlflow.autolog()
train_df = pd.DataFrame(preprocessor1.fit_transform(X_train,y_train),columns=preprocessor1.get_feature_names_out())
train_df['target'] = y_train.tolist()
summary = automl.classify(train_df, target_col="target", timeout_minutes=30)

Any ideas? Thanks.

Re: Run with UUID *** is already active when running automl

shan_chandra — Tue, 31 Jan 2023 16:08:49 GMT

@Al IDI - could you please let us know the ML runtime version you have ran into this?

could you please try setting and see if it works?

spark.conf.set("spark.databricks.mlflow.trackHyperopt.enabled", "false")

topic Re: Run with UUID *** is already active when running automl in Machine Learning

Run with UUID *** is already active when running automl

Re: Run with UUID *** is already active when running automl