import json import mlflow import pandas as pd from pprint import pprint from mlflow.tracking import MlflowClient model_name = "s434804" model_version = 12 mlflow.set_tracking_uri("http://172.17.0.1:5000") model = mlflow.keras.pyfunc.load_model( model_uri=f"models:/{model_name}/{model_version}", ) """ Honestly I have no clue how to get model's json input example any other way, so just intialize MLFlow client, get latest version of choosen model, and then get path to model's files """ client = MlflowClient() models_version = client.search_model_versions("name='s434804'") path_to_input = models_version[-1].source with open(f'{path_to_input}/input_example.json', 'r') as datafile: data = json.load(datafile) example_input = data["inputs"] input_dictionary = {i: x for i, x in enumerate(example_input)} input_ex = pd.DataFrame(input_dictionary, index=[0]) print(model.predict(input_ex))