Hugging Face for MLOps

Learn how to leverage Hugging Face and its powerful machine learning capabilities. Build, train, and deploy your own models using the Hugging Face platform and libraries.

In this course you'll get hands-on with Hugging Face and learn how to:

• Access and use pre-trained models from the Hub
• Fine-tune models with your own data
• Build machine learning pipelines with Hugging Face Transformers
• Add and version your own datasets
• Containerize and deploy Hugging Face models
• Automate workflows with GitHub Actions

By the end, you'll have practical experience building, training, and deploying Hugging Face models, including production deployment to the Azure cloud.

Learning objectives

• Find and use pre-trained models
• Fine-tune models for custom tasks
• Build ML pipelines with Hugging Face libraries
• Create and version datasets
• Containerize models for production
• Automate workflows for MLOps

Lesson 1: Getting Started with Hugging Face

Lesson Outline

• Overview of Hugging Face Hub
• Browsing models and datasets
• Using Hugging Face repositories
• Managing spaces and access

Lesson 2: Applying Hugging Face Models

Lesson Outline

• Downloading models from the Hub
• Using models with PyTorch/TensorFlow
• Leveraging tokenizers and pipelines
• Performing inference with Hub models

Lesson 3: Working with Datasets

Lesson Outline

• Browsing datasets on Hugging Face
• Uploading and managing datasets
• Versioning datasets with dataset cards
• Loading datasets in PyTorch/TensorFlow

Lesson 4: Model Serving and Deployment

Lesson Outline

• Containerizing Hugging Face models
• Creating inference APIs with FastAPI
• Deploying to cloud services like Azure
• Automating with GitHub Actions

About your instructor

Alfredo Deza has over a decade of experience as a Software Engineer doing DevOps, automation, and scalable system architecture. Before getting into technology he participated in the 2004 Olympic Games and was the first-ever World Champion in High Jump representing Peru. He currently works in Developer Relations at Microsoft and is an Adjunct Professor at Duke University. This solid background in technology and teaching, including his experience teaching and authoring content about LOps will give you everything you need to get started applying these powerful concepts.

Resources

• MLOps with Databricks

• Introduction to MLflow for MLOps

• Hands-on Python for MLOps