By default, the MLflow client saves artifacts to an artifact store URI during an experiment. The artifact store URI is similar to /dbfs/databricks/mlflow-tracking/<experiment-id>/<run-id>/artifacts/.

This artifact store is a MLflow managed location, so you cannot download artifacts directly.

You must use client.download_artifacts in the MLflow client to copy artifacts from the artifact store to another storage location.

Example code

This example code downloads the MLflow artifacts from a specific run and stores them in the location specified as local_dir.

Replace <local-path-to-store-artifacts> with the local path where you want to store the artifacts.

Replace <run-id> with the run_id of your specified MLflow run.

%python

import mlflow
import os
from mlflow.tracking import MlflowClient
client = MlflowClient()
local_dir = "<local-path-to-store-artifacts>"
if not os.path.exists(local_dir):
  os.mkdir(local_dir)

# Creating sample artifact "features.txt".
features = "rooms, zipcode, median_price, school_rating, transport"
with open("features.txt", 'w') as f:
    f.write(features)

# Creating sample MLflow run & logging artifact "features.txt" to the MLflow run.
with mlflow.start_run() as run:
    mlflow.log_artifact("features.txt", artifact_path="features")

# Download the artifact to local storage.
local_path = client.download_artifacts(<run-id>, "features", local_dir)
print("Artifacts downloaded in: {}".format(local_dir))
print("Artifacts: {}".format(local_dir))

After the artifacts have been downloaded to local storage, you can copy (or move) them to an external filesystem or a mount point using standard tools.

Copy to an external filesystem

%scala

dbutils.fs.cp(local_dir, "<filesystem://path-to-store-artifacts>")

Move to a mount point

%python

shutil.move(local_dir, "/dbfs/mnt/<path-to-store-artifacts>")