Machine Learning (ML) on AWS - ML Models and Tools - AWS

Introduction

• AWS provides a comprehensive set of services and infrastructure for machine learning (ML).
• More than 100,000 customers use AWS ML services for business solutions and innovation.
• Amazon SageMaker is a central tool for building, training, and deploying ML models.

Key Services and Tools

Amazon SageMaker

• Facilitates building, training, and deploying ML models at scale.
• Supports training and fine-tuning of foundation models quickly.
• Allows utilization of popular AI development apps within SageMaker AI.

AWS Deep Learning AMIs and Containers

• AWS Deep Learning AMIs: Preconfigured environments for secure and scalable deep learning applications.
• AWS Deep Learning Containers: Optimized, prepackaged container images for deploying deep learning environments.

Frameworks

• Hugging Face on Amazon SageMaker: Train and deploy Hugging Face models quickly.
• TensorFlow on AWS: Tools for enhancing and visualizing deep learning applications.
• PyTorch on AWS: Enterprise-ready PyTorch experience for scalable ML.
• Apache MXNet on AWS: Rapidly build and run ML applications.

AI Infrastructure

• Amazon EC2 Instances
  • Trn1: Cost-effective training for generative AI models.
  • P5: High-performance GPU-based instances for deep learning.
  • Inf2: High performance for generative AI inference.
  • G5: GPU-based instances for graphics and ML inference.
• Amazon SageMaker HyperPod
  • Infrastructure for distributed training at scale.

Customer Innovations

• Over 100,000 customers across industries use AWS ML for better service, optimization, and innovation.
• Examples:
  • Amazon Ads: Generative AI for custom image creation.
  • Perplexity: Accelerated FM training by 40% using SageMaker HyperPod.
  • Booking.com: Personalized accommodation recommendations.
  • Itau: Improved ML solutions productivity and speed to market.
  • BMW Group: Scalable ML environment using SageMaker Studio.

Responsible AI

• AWS focuses on developing AI responsibly with a people-centric approach.
• Tools include Guardrails for Amazon Bedrock and Amazon SageMaker Clarify.

Learning and Development

• AWS Solutions Library: Curated solutions for common AI use cases.
• AWS DeepRacer League: Autonomous racing league for expanding ML skills.
• Amazon SageMaker Studio Lab: Platform for learning and experimenting with ML.
• ML Tutorials: Guide on using Amazon SageMaker for ML lifecycle tasks.
• AWS ML Community: Network of AWS ML customers, influencers, and experts.
• AI Courses: Training resources for data scientists and ML engineers.

Getting Started

• Training and Certification: Build ML skills with Amazon's curriculum.
• AWS Partners: Collaborate with partners for AI innovation.

Additional Resources

• Explore more AI infrastructure capabilities
• Browse more ML customer stories