Llama 3.2 models from Meta are now available in Amazon SageMaker JumpStart

This article announces the availability of Meta's Llama 3.2 large language models (LLMs) in Amazon SageMaker JumpStart, which provides access to pre-trained models. Llama 3.2 models offer enhanced capabilities, including multi-modality support for vision tasks and lightweight models suitable for edge devices.

Specifically, the article covers:

• Overview of Llama 3.2 models, their sizes (1B, 3B, 11B, 90B), and key features like multi-modal vision support, context length, and lightweight models for edge devices
• Introduction to SageMaker JumpStart and its benefits for deploying pre-trained models
• Step-by-step guide on discovering and deploying Llama 3.2 models, specifically the 11B Vision model, using SageMaker JumpStart
• List of recommended instances and supported context lengths for various Llama 3.2 models
• Examples of using Llama 3.2 models for text-only, single-image, and multi-image input tasks
• Information on cleaning up deployed resources