I'm attempting to use MLFlow to register models in Databricks and am following the recipe at...

https://nlp.johnsnowlabs.com/docs/en/licensed_serving_spark_nlp_via_api_databricks_mlflow

when i execute...

mlflow.spark.log_model(pipeline, "lemmatizer", conda_env=conda_env)

i get the following error....

MlflowException: Cannot serialize this model. MLflow can only save descendants of pyspark.Modelthat implement MLWritable and MLReadable.

can anyone help?