Getting Started with MLflow
MLflow 入门指南

For those new to MLflow or seeking a refresher on its core functionalities, the quickstart tutorials here are the perfect starting point. They will guide you step-by-step through fundamental concepts, focusing purely on a task that will maximize your understanding of how to use MLflow to solve a particular task.
对于刚接触 MLflow 或希望复习其核心功能的用户,这里的快速入门教程是理想的起点。它们将引导您一步步了解基本概念,专注于一个能最大化理解如何使用 MLflow 解决特定任务的内容。

Guidance on Running Tutorials
教程运行指南

If you are new to MLflow and have never interfaced with the MLflow Tracking Server, we highly encourage you to head on over to quickly read the guide below. It will help you get started as quickly as possible with tutorial content throughout the documentation.
如果您是 MLflow 的新手,并且从未与MLflow 跟踪服务器进行过交互,我们强烈建议您立即前往阅读下面的指南。它将帮助您通过文档中的教程内容,尽可能快速地入门。

Getting Started Guides 入门指南

MLflow Tracking MLflow 跟踪

MLflow Tracking is one of the primary service components of MLflow. In these guides, you will gain an understanding of what MLflow Tracking can do to enhance your MLOps related activities while building ML models.
MLflow 跟踪 是 MLflow 的主要服务组件之一。在这些指南中,您将了解 MLflow 跟踪如何增强您在构建 ML 模型时的 MLOps 相关活动。

The basics of MLflow tracking.

In these introductory guides to MLflow Tracking, you will learn how to leverage MLflow to:
在这些 MLflow 跟踪的入门指南中,您将学习如何利用 MLflow 来:

  • Log training statistics (loss, accuracy, etc.) and hyperparameters for a model
    日志模型训练统计数据(损失、准确率等)和超参数

  • Log (save) a model for later retrieval
    日志(保存)模型以供日后检索

  • Register a model using the MLflow Model Registry to enable deployment
    注册一个模型使用MLflow 模型注册表以启用部署

  • Load the model and use it for inference
    加载模型并用于推理

In the process of learning these key concepts, you will be exposed to the MLflow Tracking APIs, the MLflow Tracking UI, and learn how to add metadata associated with a model training event to an MLflow run.
在学习这些关键概念的过程中,您将接触到MLflow Tracking APIs、MLflow Tracking UI,并学习如何将与模型训练事件相关的元数据添加到 MLflow 运行中。

Autologging Basics 自动日志记录基础

A great way to get started with MLflow is to use the autologging feature. Autologging automatically logs your model, metrics, examples, signature, and parameters with only a single line of code for many of the most popular ML libraries in the Python ecosystem.
开始使用 MLflow 的一个绝佳方式是利用 autologging 功能。Autologging 能够自动记录您的模型、指标、示例、签名及参数,仅需一行代码即可支持 Python 生态系统中许多最流行的机器学习库。

The basics of MLflow tracking.

In this brief tutorial, you’ll learn how to leverage MLflow’s autologging feature to simplify your model logging activities.
在本简短教程中,您将学习如何利用 MLflow 的自动记录功能来简化您的模型记录活动。

Run Comparison Basics 运行比较基础

This quickstart tutorial focuses on the MLflow UI’s run comparison feature and provides a step-by-step walkthrough of registering the best model found from a hyperparameter tuning execution. After locally serving the registered model, a brief example of preparing a model for remote deployment by containerizing the model using Docker is covered.
本快速入门教程重点介绍 MLflow UI 的运行比较功能,并提供从超参数调优执行中注册最佳模型的分步指南。在本地服务注册模型后,简要介绍了通过使用 Docker 容器化模型来准备模型进行远程部署的示例。

The basics of MLflow run comparison.

Tracking Server Quickstart
跟踪服务器快速入门

This quickstart tutorial walks through different types of MLflow Tracking Servers and how to use them to log your MLflow experiments.
本快速入门教程介绍了不同类型的MLflow 跟踪服务器,并展示了如何使用它们记录您的 MLflow 实验。

Model Registry Quickstart
模型注册表快速入门

This quickstart tutorial walks through registering a model in the MLflow model registry and how to retrieve registered models.
本快速入门教程将引导您完成在 MLflow 模型注册表中注册模型以及如何检索已注册模型的过程。

Further Learning - What’s Next?
深入学习 - 下一步是什么?

Now that you have the essentials under your belt, below are some recommended collections of tutorial and guide content that will help to broaden your understanding of MLflow and its APIs.
既然您已经掌握了基础知识,以下是一些推荐的教程和指南内容集合,将有助于拓宽您对 MLflow 及其 API 的理解。

  • Tracking - Learn more abou the MLflow tracking APIs by reading the tracking guide.
    追踪 - 通过阅读追踪指南了解更多关于 MLflow 追踪 API 的信息。

  • LLMs - Discover how you can leverage cutting-edge advanced LLMs to power your ML applications by reading the LLMs guide. 

  • MLflow Deployment - Follow the comprehensive guide on model deployment to learn how to deploy your MLflow models to a variety of deployment targets. 

  • Model Registry - Learn about the MLflow Model Registry and how it can help you manage the lifecycle of your ML models. 

  • Deep Learning Library Integrations - From PyTorch to TensorFlow and more, learn about the integrated deep learning capabilities in MLflow by reading the deep learning guide. 

  • Traditional ML - Learn about the traditional ML capabilities in MLflow and how they can help you manage your traditional ML workflows.