MLFlow Error

[adithyasanker]

On running https://www.databricks.com/notebooks/mlops/deploy_azure_ml_model_.html

At the build docker image step, 

mlflow.azureml.build_image

I faced an error ''TypeError: join() argument must be str, bytes, or os.PathLike object, not 'dict''' .

The model is now registered successfully in azureml but the docker image creation fails. I'm using the Spark 12.2 LTS ML Compute.

Could this be an mlflow version issue?

[Kaniz]

Hi @adithyasanker, This error is typically related to Python and occurs when the os.path.join() function is provided with a dictionary as an argument, which is not a valid input type. As for your question about whether this could be an MLflow version issue, it's possible, but I'm not sure. MLflow interacts with Docker during the model serving process, and a mismatch between the MLflow version and the Docker or Databricks runtime version could cause issues. I recommend checking the code you're using os.path.join() to ensure you're not providing a dictionary as an argument. If you're using MLflow, you might also want to verify that your MLflow version is compatible with your Docker and Databricks runtime versions.