Highlights from re:Invent Day 3 include discussions on Amazon Bedrock, Amazon Titan text models, Drop xCloud 2.1, Meadows Llama 70 Billion Parameter, Stable Diffusion Excel 1.0 image model, smart sifting with Amazon SageMaker, new features in Amazon SageMaker Canvas, Open Search Service, and MemoryDB for Redis, and updates from AWS including Vector Search in Amazon DocumentDB, AWS Cleanrooms ML, healthcare modeling in AWS Cleanrooms, Apache Iceberg support in Amazon Redshift, Neptune Analytics for graph data, and new AWS AI service cards.
Amazon Bedrock now supports batch inference and model evaluation, providing users with customizable evaluation metrics.
SageMaker HyperPod and smart sifting capability in Amazon SageMaker optimize distributed training and reduce training time and cost.
Deep dives
Amazon Bedrock Updates
Amazon Bedrock now supports batch inference and model evaluation. It offers automatic and human evaluation workflows, allowing users to customize evaluation metrics. The Amazon Titan multimodal embeddings foundation model is now accessible in Amazon Bedrock, enabling accurate and contextually relevant multimodal search and recommendation experiences.
Amazon Titan Text Models in Bedrock
Amazon Titan text models Express and Light are now generally available in Amazon Bedrock. These language models assist in various text-related tasks, offering options for different use cases and performance needs.
Updates to SageMaker and Open Search
The podcast episode features several updates related to SageMaker and Open Search. Some notable updates include the introduction of SageMaker HyperPod for distributed training at scale, the preview of smart sifting capability in Amazon SageMaker that reduces training time and cost, and the introduction of OR1 instances in Open Search, optimized for indexing heavy workloads. Additionally, Amazon SageMaker Clarify now supports foundation model evaluations, and Amazon SageMaker Canvas expands capabilities for enhancing foundation models' performance. There are also updates for Amazon SageMaker pipelines, inference capabilities, and the introduction of Amazon MemoryDB for Redis, supporting vector search. Finally, AWS Cleanrooms ML is now available in preview, enabling privacy-enhancing machine learning collaborations without sharing raw data.
Our re:Invent 2023 coverage continues! In this episode, Jillian Forde shares highlights from re:Invent Day 3, focusing on the profound and impactful relationship between humans, data, and AI that is unfolding right before us.
Get the Snipd podcast app
Unlock the knowledge in podcasts with the podcast player of the future.
AI-powered podcast player
Listen to all your favourite podcasts with AI-powered features
Discover highlights
Listen to the best highlights from the podcasts you love and dive into the full episode
Save any moment
Hear something you like? Tap your headphones to save it with AI-generated key takeaways
Share & Export
Send highlights to Twitter, WhatsApp or export them to Notion, Readwise & more
AI-powered podcast player
Listen to all your favourite podcasts with AI-powered features
Discover highlights
Listen to the best highlights from the podcasts you love and dive into the full episode