Nebari Plugin MLflow AWS
Table of Contents
Introduction
This MLflow extension is designed to integrate into Nebari deployments utilizing the AWS provider. It provides a robust, collaborative environment for AI/ML professionals to manage experiments, track metrics, and deploy models.
Features
Centralized Artifact Repository: Store and manage all your metrics, parameters, and artifacts in a single location, accessible across the multi-tenant platform.
Experiment Tracking: Log, query, and visualize metrics to understand and compare different runs and models.
Automated Configuration: Simply type import mlflow in your Python script, and you're already configured to communicate with the remote multi-tenant MLflow server—no additional setup required.
Installation
Prerequisites:
- Nebari must be deployed using the provider AWS
- Nebari version 2024.5.1 or later
Installing the MLflow extension is as straightforward as installing a Python package. Run the following commands:
git clone nebari-plugin-mlflow-aws
cd nebari-plugin-mlflow-aws/
pip install nebari-plugin-mlflow-aws
This command installs the Python package and also creates the necessary infrastructure to run MLflow on the AI Platform.
Configuration
After installation, the MLflow extension is automatically configured to work with the AI Platform. To access the MLflow interface, navigate to https://[your-nebari-domain]/mlflow.
Configuring MLFlow Kubernetes Namespace
In order for MLflow to deploy into a non-default Kubernetes namespace, you may add an mlflow block to your Nebari configuration file such as the example below. Although MLflow resources should all be prefixed to avoid collisions, we still recommend this as a best practice.
mlflow:
namespace: mlflow
Exposing MLflow workload via Ingress
In order for the traefik ingress to route users' web requests to the MLflow workload, add or update the block in your Nebari configuration file. Be sure to update this block if you've configured a namespace other than mlflow.
ingress:
terraform_overrides:
additional-arguments:
- "--providers.kubernetescrd.namespaces=mlflow"
Configuring MLflow Tracking URL
You may set the MLFLOW_TRACKING_URL to configure mlflow in individual users' Nebari instances by adding or updating an additional block in your Nebari configuration file. Note that the first 'mlflow' in the URL below corresponds to the Kubernetes namespace where mlflow is deployed, so if you have assigned a custom namespace, that URL will need updating.
jupyterhub:
overrides:
singleuser:
extraEnv:
MLFLOW_TRACKING_URI: "http://mlflow.mlflow:5000"
Usage
Getting started with the MLflow extension is incredibly simple. To track an experiment:
Navigate to the MLFLow extension URL and create a new experiment. In your Python script, import MLflow and start logging metrics.
import mlflow
# Start an experiment
with mlflow.start_run() as run:
mlflow.log_metric("accuracy", 0.9)
mlflow.log_artifact("path/to/your/artifact")
With the above code, your metrics and artifacts are automatically stored and accessible via the MLFlow extension URL.
License
nebari-plugin-mlflow-aws is distributed under the terms of the Apache license.
Metadata
Release files for nebari-plugin-mlflow-aws 0.0.16
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| nebari_plugin_mlflow_aws-0.0.16.tar.gz | 22.3 kB | Without provenance | Details for nebari_plugin_mlflow_aws-0.0.16.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| nebari_plugin_mlflow_aws-0.0.16-py3-none-any.whl | Python 3 | none | any | Details for nebari_plugin_mlflow_aws-0.0.16-py3-none-any.whl |
Total release size: 49.9 kB
Release files / nebari_plugin_mlflow_aws-0.0.16.tar.gz
| Download URL | nebari_plugin_mlflow_aws-0.0.16.tar.gz |
|---|---|
| Size | 22.3 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
3d2bd12cde5951a0f1b768c6c53a0513531d058924eb206643a04c1c5dfb050b
|
|
BLAKE2b-256 checksum How to use checksums |
875d943fab220f1a2ba98601cc71c4150f78ae007e7499414e5d57f6b0fcfa77
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/5.1.0 CPython/3.12.4
|
Release files / nebari_plugin_mlflow_aws-0.0.16-py3-none-any.whl
| Download URL | nebari_plugin_mlflow_aws-0.0.16-py3-none-any.whl |
|---|---|
| Size | 27.5 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
4f853f1536aad9759628138440dcc79ba414783e9a086bad0a4ebf8574a60af2
|
|
BLAKE2b-256 checksum How to use checksums |
81b64f4a3b0a630fc2b18aa4593f3b6cd88b5cdd44ee6d6f07eae3e6e1ef886b
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/5.1.0 CPython/3.12.4
|
Security for nebari-plugin-mlflow-aws 0.0.16
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.