Skip to content

Latest commit

 

History

History
391 lines (316 loc) · 12.8 KB

README_single.md

File metadata and controls

391 lines (316 loc) · 12.8 KB
Raw

MLflow Export Import - Single Object Tools

Overview

The Single tools allow you to export and import individual MLflow objects between tracking servers. These tools allow you to specify a different destination object name.

For example, if you want to clone the experiment /Mary/Experiment/Iris under a new name, you can specify the target experiment name as /John/Experiment/Iris.

See sample JSON export files here.

Tools

MLflow Object Documentation Code
Model export-model code
import-model code
Experiment export-experiment code
import-experiment code
Run export-run code
import-run code

Experiment Tools

Export Experiment

Export an experiment to a directory. Accepts either an experiment ID or name.

Usage

export-experiment --help

Options:
  --experiment TEXT             Experiment name or ID.  [required]
  --output-dir TEXT             Output directory.  [required]
  --export-permissions BOOLEAN  Export Databricks permissions.  [default:
                                False]
  --run-start-time TEXT         Only export runs started after this UTC time
                                (inclusive). Format: YYYY-MM-DD.
  --notebook-formats TEXT       Databricks notebook formats. Values are
                                SOURCE, HTML, JUPYTER or DBC (comma
                                seperated).

Examples

Export experiment by experiment ID.

export-experiment \
  --experiment 2 \
  --output-dir out

Export experiment by experiment name.

export-experiment \
  --experiment sklearn-wine \
  --output-dir out

Databricks export examples

See Access the MLflow tracking server from outside Databricks.

export MLFLOW_TRACKING_URI=databricks
export DATABRICKS_HOST=https://mycompany.cloud.databricks.com
export DATABRICKS_TOKEN=MY_TOKEN

export-experiment \
  --experiment /Users/[email protected]/SklearnWine \
  --output-dir out \
  --notebook-formats DBC,SOURCE

Export directory structure

The export directory contains a JSON export file for the experiment and a subdirectory for each run. The run directory contains a JSON export file containing run metadata and an artifact folder directory.

Sample export directory

+-experiment.json
+-eb66c160957d4a28b11d3f1b968df9cd/
| +-run.json
| +-artifacts/
|   +-plot.png
|   +-model/
|     +-requirements.txt
|     +-python_env.yaml
|     +-model.pkl
|     +-conda.yaml
|     +-MLmodel

Import Experiment

Import an experiment from a directory. Reads the manifest file to import the expirement and its runs.

The experiment will be created if it does not exist in the destination tracking server. If the destination experiment already exists, the source runs will be added to it.

Usage

import-experiment --help 

Options:
  --experiment-name TEXT        Destination experiment name  [required]
  --input-dir TEXT              Input path - directory  [required]
  --import-source-tags BOOLEAN  Import source information for registered model
                                and its versions ad tags in destination
                                object.  [default: False]
  --use-src-user-id BOOLEAN     Set the destination user field to the source
                                user field.  Only valid for open source
                                MLflow.  When importing into Databricks, the
                                source user field is ignored since it is
                                automatically picked up from your Databricks
                                access token.  There is no MLflow API endpoint
                                to explicity set the user field for any
                                objects such as Run or Experiment.
  --dst-notebook-dir TEXT       Databricks destination workpsace base
                                directory for notebook. A run ID will be added
                                to contain the run's notebook.

Import examples

import-experiment \
  --experiment-name imported_sklearn \
  --input-dir out

Databricks import examples

export MLFLOW_TRACKING_URI=databricks
import-experiment \
  --experiment-name /Users/[email protected]/imported/SklearnWine \
  --input-dir exported_experiments/3532228 \

Run Tools

Export run

Export run to directory.

Usage

export-run --help

Options:
  --run-id TEXT                   Run ID.  [required]
  --output-dir TEXT               Output directory.  [required]
  --notebook-formats TEXT         Notebook formats. Values are SOURCE, HTML,
                                  JUPYTER or DBC (comma seperated).  [default: ]

Run examples

export-run \
  --run-id 50fa90e751eb4b3f9ba9cef0efe8ea30 \
  --output-dir out

Produces a directory with the following structure:

+-run.json
+-artifacts/
| +-model/
|   +-requirements.txt
|   +-python_env.yaml
|   +-model.pkl
|   +-conda.yaml
|   +-MLmodel

Sample run.json files: OSS - Databricks.

Import run

Imports a run from a directory.

Usage

import-run --help

Options:
  --input-dir TEXT                Input path - directory  [required]
  --import-source-tags BOOLEAN    Import source information for registered
                                  model and its versions ad tags in
                                  destination object.  [default: False]
  --experiment-name TEXT          Destination experiment name  [required]
  --use-src-user-id BOOLEAN       Set the destination user field to the source
                                  user field.  Only valid for open source
                                  MLflow.  When importing into Databricks, the
                                  source user field is ignored since it is
                                  automatically picked up from your Databricks
                                  access token.  There is no MLflow API
                                  endpoint to explicity set the user field for
                                  any objects such as Run or Experiment. XX
  --dst-notebook-dir TEXT         Databricks destination workpsace base
                                  directory for notebook. A run ID will be
                                  added to contain the run's notebook.
  --mlmodel-fix BOOLEAN           Add correct run ID in destination MLmodel
                                  artifact. Can be expensive for deeply nested
                                  artifacts.  [default: True]
  --dst-notebook-dir-add-run-id TEXT
                                  Add the run ID to the destination notebook
                                  workspace directory.

Import examples

Local import example
export MLFLOW_TRACKING_URI=databricks

import-run \
  --run-id 50fa90e751eb4b3f9ba9cef0efe8ea30 \
  --input out \
  --experiment-name sklearn_wine_imported
Databricks import example
export MLFLOW_TRACKING_URI=databricks

run.import-run \
  --run-id 50fa90e751eb4b3f9ba9cef0efe8ea30 \
  --input out \
  --experiment-name /Users/[email protected]/imported/SklearnWine

Registered Model Tools

Export Registered Model

Export a registered model to a directory. The default is to export all versions of a model including all None and Archived stages. You can specify a list of stages to export.

Source: export_model.py.

Usage

export-model --help

Options:
  --model TEXT                    Registered model name.  [required]
  --output-dir TEXT               Output directory.  [required]
  --notebook-formats TEXT         Databricks notebook formats. Values are
                                  SOURCE, HTML, JUPYTER or DBC (comma
                                  seperated).
  --stages TEXT                   Stages to export (comma seperated). Default
                                  is all stages and all versions. Stages are
                                  Production, Staging, Archived and None.
                                  Mututally exclusive with option --versions.
  --versions TEXT                 Export specified versions (comma separated).
                                  Mututally exclusive with option --stages.
  --export-latest-versions BOOLEAN
                                  Export latest registered model versions
                                  instead of all versions.  [default: False]
  --export-permissions BOOLEAN    Export Databricks permissions.  [default:
                                  False]
  --get-model-version-download-uri BOOLEAN
                                  Call MLflowClient.get_model_version_download
                                  _uri() for version export.  [default: False]

Example

export-model \
  --model sklearn_wine \
  --output-dir out \
  --stages Production,Staging

Found 6 versions
Exporting version 3 stage 'Production' with run_id 24aa9cce1388474e9f26d17100724cdd to out/24aa9cce1388474e9f26d17100724cdd
Exporting version 5 stage 'Staging' with run_id 8efd80f59b7946119d8f1838515eea25 to out/8efd80f59b7946119d8f1838515eea25

Output

Output export directory example.

+-749930c36dee49b8aeb45ee9cdfe1abb/
| +-artifacts/
|   +-sklearn-model/
|   | +-model.pkl
|   | +-conda.yaml
|   | +-MLmodel
|   |  
+-model.json

Sample model.json files: OSS - Databricks.

{
"mlflow": {
  "registered_model": {
    "name": "sklearn_wine",
    "creation_timestamp": "1587517284168",
    "last_updated_timestamp": "1587572072601",
    "description": "hi my desc",
    "latest_versions": [
      {
        "name": "sklearn_wine",
        "version": "1",
        "creation_timestamp": "1587517284216",
. . .

Import registered model

Import a registered model from a directory.

Source: import_model.py.

Usage

import-model --help

Options:

--input-dir TEXT Input directory [required] --model TEXT Registered model name. [required] --experiment-name TEXT Destination experiment name [required] --delete-model BOOLEAN If the model exists, first delete the model and all its versions. [default: False] --import-source-tags BOOLEAN Import source information for registered model and its versions ad tags in destination object. [default: False] --await-creation-for INTEGER Await creation for specified seconds. --sleep-time INTEGER Sleep time for polling until version.status==READY. --verbose BOOLEAN Verbose. [default: False]


#### Example

import-model
--model sklearn_wine
--experiment-name sklearn_wine_imported
--input-dir out
--delete-model True

Model to import: Name: sklearn_wine Description: my model 2 versions Deleting 1 versions for model 'sklearn_wine_imported' version=2 status=READY stage=Production run-id=f93d5e4d182e4f0aba5493a0fa8d9eb6 Importing versions: Version 1: current_stage: None: Run to import: run-id: 749930c36dee49b8aeb45ee9cdfe1abb artifact_uri: file:///opt/mlflow/server/mlruns/1/749930c36dee49b8aeb45ee9cdfe1abb/artifacts source: file:///opt/mlflow/server/mlruns/1/749930c36dee49b8aeb45ee9cdfe1abb/artifacts/sklearn-model model_path: sklearn-model run-id: 749930c36dee49b8aeb45ee9cdfe1abb Importing run into experiment 'scratch' from 'out/749930c36dee49b8aeb45ee9cdfe1abb' Imported run: run-id: 03d0cfae60774ec99f949c42e1575532 artifact_uri: file:///opt/mlflow/server/mlruns/13/03d0cfae60774ec99f949c42e1575532/artifacts source: file:///opt/mlflow/server/mlruns/13/03d0cfae60774ec99f949c42e1575532/artifacts/sklearn-model Version: id=1 status=READY state=None Waited 0.01 seconds