From Hugging Face to Amazon SageMaker Studio in one click

Hugging Face has announced a deep-link integration with Amazon SageMaker AI that allows developers to transition from model discovery to hands-on experimentation in SageMaker Studio with a single selection. Previously, setting up SageMaker Studio after discovering a model on Hugging Face involved multiple manual steps, such as navigating the AWS Console, creating domains, configuring IAM permissions, and checking GPU quotas. According to the announcement, the new launch introduces capabilities designed to shorten this path. When browsing supported models on Hugging Face, users can select actions like Customize on SageMaker AI or Deploy on SageMaker AI to land directly in the console with model context preserved. SageMaker AI automatically provisions a new domain with pre-configured permissions, including a newly created managed policy called AmazonSageMakerModelCustomizationCoreAccess. This policy handles serverless model customization jobs using methods such as supervised fine-tuning, direct preference optimization, and reinforcement learning, eliminating the need to manually configure IAM roles. Additionally, the Studio UI now displays GPU instance quota availability directly in the instance selection list, allowing developers to immediately view account limits or be redirected to Service Quotas if a limit increase is required. Arcee noted that the integration helps take open models the last mile by letting users go straight from Hugging Face into SageMaker Studio to fine-tune or deploy within their own AWS environments. To get started, users can visit the Amazon SageMaker Studio page or explore models on Hugging Face.