Introducing Storage Buckets

Hugging Face Buckets are a new storage primitive on the Hugging Face Hub for mutable ML artifacts like checkpoints, traces, logs, and pipeline outputs. In this video I explain where Buckets come from, why Xet deduplication matters, how they compare to Git-based repositories, and how to create and use them from the GUI, CLI, and Python. If you work with training loops, agents, or repeated experiment outputs, this is the practical storage workflow to know. --- 🤓 *Topics Covered* - Hugging Face Buckets overview - Xet deduplication for ML artifacts - CLI and Python bucket workflows --- ⏱️ *Timestamps* 0:00 Introduction 0:56 What Hugging Face Buckets are 5:35 Bucket creation demo --- ## Connect with me - X: https://x.com/_alejandroao - LinkedIn: https://www.linkedin.com/in/alejandro-ao/