[Python] mlflow cannot fetch model registered on GitLab model registry

I’m trying to download artifacts and a model that has been saved using MLflow on GitLab model registry. That's the working part of my code, where I set up the client, create model and version and upload artifact previously saved on a local folder:

# Running client
import os
from mlflow import MlflowClient

os.environ["MLFLOW_TRACKING_URI"] = "..."
os.environ["MLFLOW_TRACKING_TOKEN"] = "..."

client = MlflowClient()

# Model and version creation
model_name = 'test_model_n'

description = 'test model'
model = client.create_registered_model(model_name, description=description)

model_version = '1.0.1'
tags = { "gitlab.version": model_version }
model_version = client.create_model_version(model_name, model_version, tags=tags)

# Logging artifacts

run_id = model_version.run_id
client.log_artifact(run_id, 'mlartifacts/875771393960734527', artifact_path="")



Now, in order to fetch the model i tried multiple approaches:

Method 1


model_uri = f"models:/{model_name}/{model_version}"
model = mlflow.pyfunc.load_model(model_uri)



Download bar reach 100%, but a warning and an error pops out:

WARNING:urllib3.connectionpool:Connection pool is full, discarding connection: "gitlab server" Connection pool size: 10
Could not find a registered artifact repository for: c:. Currently registered schemes are: ['', 'file', 's3', 'r2', 'gs', 'wasbs', 'ftp', 'sftp', 'dbfs', 'hdfs', 'viewfs', 'runs', 'models', 'http', 'https', 'mlflow-artifacts', 'abfss']

Method 2


artifacts_path, = mlflow.artifacts.download_artifacts(
run_id=run_id,
artifact_path="",
dst_path="./downloaded_model"



Error:

ValueError: not enough values to unpack (expected 2, got 1)

Method 3



# Se conosci il run_id specifico
mv = client.get_model_version(model_name, model_version)
run_id = mv.run_id

# Costruisci l'URI usando il run_id
run_uri = f"runs:/{run_id}/[path_to_the_model]"

model_from_run = mlflow.pyfunc.load_model(run_uri)


Error:

ValueError: not enough values to unpack (expected 2, got 1)


Any help would be very appreciated, feel free to ask for more info if you think it would help.

Continue reading...