Arun Kumar Lokanatha

Amazon SageMaker HyperPod accelerates FM development through purpose-built infrastructure. In this session, learn how to leverage HyperPod recipes for efficient LLM customization and use inference operators to deploy models on EKS clusters.

Learn how to customize foundation models for your specific business needs. In this chalk talk, we'll discuss various customization approaches including supervised fine-tuning, reinforcement learning, and explore when to apply each technique most effectively.

In this builders’ session, get step-by-step implementation experience fine-tuning Llama 3 8B using the QLoRA strategy. Learn how you can fine-tune Llama 3 models at scale using the memory-efficient PyTorch Fully Sharded Data Parallel (FSDP2) library with access to built-in debugging and GPU utilization metrics.

Amazon SageMaker offers purpose-built tools to reduce the time and cost to train and tune machine learning (ML) models at scale without the need to manage infrastructure. Join this chalk talk to find the right tools for your use case.