Posted On: Jul 28, 2023

Amazon SageMaker Studio launched support for the ml.G5 GPU instance family in the South America region (São Paulo). Customers using ml.G5 instances in Amazon SageMaker Studio Notebooks can leverage up to 3.3x faster performance for machine learning training compared to ml.G4dn instances. This makes them a high performance and cost-efficient solution for interactive model training for use cases such as generative AI foundation model tuning, natural language processing, computer vision, and recommender engines.

ml.G5 instances feature up to 8 NVIDIA A10G Tensor Core GPUs and 2nd generation AMD EPYC processors. They also support up to 192 vCPUs, up to 100 Gbps of network bandwidth, and up to 7.6 TB of local NVMe SSD storage. Customers can choose the most appropriate instance size from eight options, offering between one and eight GPUs. To read more about ml.G5 instances, visit the AWS news blog or visit the G5 instance page.

To set up and get started with SageMaker Studio, visit our developer guide . You can also see an example for fine tuning a Large Language Model using ml.G5 instance in SageMaker Studio in this blog post.