Import mlflow Error

[tanjil]

Hello, 

I am trying to replicate this motebook in my environment: 

mlflow-end-to-end-example - Databricks

However, I am getting the following error when I run "import mlflow": 

"TypeError: bases must be types"

How can I solve this issue? 

Thank you, 

Tanjil 

[Walter_C]

Can you share the specific cell of the notebook where you are receiving this error? Have you modified the code or it is the same? Do you have any particular libraries installed on the cluster you are using for the testing?

[tanjil]

The error is generated in cell 21, in the line "import mlflow".

No modifications have been made to the code.
Cluster details: 10.4 LTS (includes Apache Spark 3.2.1, Scala 2.12) 

[Kumaran]

Hello @tanjil 

Thank you for contacting databricks community support. Could you check what version of protobuf you have?

If you are using 10.4 ML cluster, the MLflow 1.x is not compatible with protobuf 4.x. The default version of protobuf in MLR 10.4 is 3.17.2. We recommend that you can either downgrade protobuf to 3.17.2 or upgrade MLflow to 2.x. Thanks