Azure Databricks model serving mlflow error version

Question

Hello or good evening everyone !

I recently wanted to use azure's registered models. I attached my experiment to my model and launched the serving, so far so good !

The problem comes from the requests we are trying to make to this model:

Unrecognized content type parameters: format. IMPORTANT: The MLflow Model scoring protocol has changed in MLflow version 2.0. If you are seeing this error, you are likely using an outdated scoring request format. To resolve the error, either update your request format or adjust your MLflow Model's requirements file to specify an older version of MLflow (for example, change the 'mlflow' requirement specifier to 'mlflow==1.30.0'). If you are making a request using the MLflow client (e.g. via mlflow.pyfunc.spark_udf()), upgrade your MLflow client to a version >= 2.0 in order to use the new request format. For more information about the updated MLflow Model scoring protocol in MLflow 2.0, see https://mlflow.org/docs/latest/models.html#deploy-mlflow-models.

Unable to find the real source of this problem. We have updated the version of Mlflow (from 1.30 to 2.0.1) and reused the exact format provided for sending data.

In addition, we find very few cases in this genre after several searches. The azure doc keeps coming back without finding other users in need.

Have you encountered this problem before ? How did you manage to debug it ?

thank you in advance for your help !

Answer 1

What worked for me was:

model_name = "model_name"
model_uri = f"models:/{model_name}/Production" 
mlflow.pyfunc.get_model_dependencies(model_uri, format='pip')

The last line gives you the dependencies you can add for install in the beginning of your notebook. Then add env_manager='conda' to the mlflow.pyfunc.spark_udf call