MLflow:一个用于管理机器学习生命周期的工具
MLflow 是一个开源平台,专为协助机器学习从业者和团队处理机器学习过程的复杂性而构建。MLflow 专注于机器学习项目的全生命周期,确保每个阶段都是可管理的、可追溯的和可复现的。
MLflow 入门资源
如果你是第一次探索 MLflow,这里的教程和指南是一个很好的起点。每个教程的重点都是让你尽可能快地掌握基本功能、术语、API 以及使用 MLflow 的一般最佳实践,以便在特定领域的指南和教程中增强你的学习。
Learn about the core components of MLflow
Quickstarts
Get Started with MLflow in our 5-minute tutorial
Guides
Learn the core components of MLflow with this in-depth guide to Tracking
![Core Components](_static/images/intro/learn-core-components.png)
Learn how to perform common tasks in MLflow
Guides
Autologging tutorial for effortless model tracking
Model Signatures and type validation in MLflow
Hyperparameter tuning with MLflow
![Model Development Lifecycle](_static/images/intro/model-dev-lifecycle.png)
Get started with MLflow's GenAI integrations
Quickstarts
Transformers Text Generation
LangChain Introductory Tutorial
Sentence Transformers Basic Embedding Tutorial
OpenAI Quickstart Tutorial
![GenAI with MLflow](_static/images/intro/genai-integrations.png)
Get started with MLflow's Deep Learning Library integrations
![Deep Learning with MLflow](_static/images/intro/mlflow-deep-learning.png)
GenAI 和 MLflow
探索 MLflow 中全面的 GenAI 支持。从 MLflow 部署 GenAI 模型 到 提示工程 UI 以及原生的 GenAI 聚焦的 MLflow 风格,如 open-ai、transformers 和 sentence-transformers,这里的教程和指南将帮助您开始利用这些强大模型、服务和应用程序的优势。您将学习到 MLflow 如何简化使用 GenAI 模型和开发利用它们的解决方案。重要的任务,如提示开发、提示评估、基础模型比较、微调、日志记录和部署生产级推理服务器,都由 MLflow 涵盖。
探索下面的指南和教程,开始你的旅程!
Explore the Prompt Engineering UI
Quickstarts
Learn how to use the Prompt Engineering UI
![Prompt Engineering UI](_static/images/intro/prompt-engineering-ui.png)
Learn about managed access to GenAI services with the MLflow AI Gateway
![MLflow AI Gateway](_static/images/intro/deployments-server.png)
Learn about GenAI Evaluation
Guides
Learn how to evaluate your GenAI applications with MLflow
Discover how to use MLflow Evaluate with the Prompt Engineering UI
![MLflow GenAI Evaluation](_static/images/intro/evaluate.png)
Learn about using Retrieval Augmented Generation (RAG) with MLflow
![MLflow RAG](_static/images/intro/rag.png)