Error: 'PyFuncModel' object has no attribute 'serve' , when deploy MLflow Model as a Local Inference Server

[SeanWebAi]

Has anyone else encountered the error: 'PyFuncModel' object has no attribute 'serve', when trying to deploy a model using Python?

I followed the MLflow guide https://mlflow.org/docs/2.12.1/deployment/deploy-model-locally.html and used the code below:

import mlflow
import os

mlflow.set_tracking_uri("https://mlflow.abc.com/")
os.environ["MLFLOW_TRACKING_USERNAME"] = ""
os.environ["MLFLOW_TRACKING_PASSWORD"] = ""

model = mlflow.pyfunc.load_model("runs://model")
model.serve(port=5000)

However, I always encounter the following error. As this is a straightforward command, am I missing anything?

Exception has occurred: AttributeError 'PyFuncModel' object has no attribute 'serve'.

Here is my requirements.txt for the reference.

cloudpickle==3.0.0
numpy==1.26.4
packaging==24.0
psutil==5.9.8
pyyaml==6.0.1
scikit-learn==1.4.2
scipy==1.13.0
virtualenv
mlflow[extras]
mlserver
mlserver-mlflow