SageMaker のアップデート

<p>AWS Glue Data Quality now supports a new Distribution Analyzer that generates frequency distribution profiles for your data. Using this new Distribution Analyzer in the Data Quality Definition Lang

<p>We are pleased to announce the availability of <a href="https://aws.amazon.com/ec2/instance-types/g7e/">Amazon EC2 G7e instances </a>in Asia Pacific (Seoul), Europe (London), and Asia Pacific (Toky

<p>We are pleased to announce the availability of <a href="https://aws.amazon.com/ec2/instance-types/g6/">Amazon EC2 G6 </a>instances in the AWS GovCloud (US-East) region on Amazon SageMaker AI infere

<p>Amazon SageMaker AI inference now supports <a href="https://aws.amazon.com/ec2/instance-types/g7/">G7 instances</a> powered by NVIDIA RTX PRO 4500 Blackwell Server Edition GPUs, enabling you to dep

<p>Amazon SageMaker Unified Studio now supports Amazon OpenSearch as a data source, enabling you to query and analyze your search and log analytics data directly alongside your other data assets. With

<p>Amazon SageMaker HyperPod now supports network topology configuration at the partition level for Slurm orchestrated clusters. A single cluster can now run tree topology in one partition and block t

<p>Today, AWS announced the availability of gemma-4-E2B-it in Amazon SageMaker JumpStart, expanding the portfolio of foundation models available to AWS customers. This model from Google DeepMind is a

<p>Today, AWS announced the availability of privacy-filter in Amazon SageMaker JumpStart, expanding the portfolio of foundation models available to AWS customers. This model from OpenAI is a bidirecti

<p>Today, AWS announced the availability of Qwen3-VL-Embedding-2B and Qwen3-Reranker-4B in Amazon SageMaker JumpStart, expanding the portfolio of foundation models available to AWS customers. These mo

<p>Today, AWS announced the availability of Voxtral-Mini-4B-Realtime-2602 in Amazon SageMaker JumpStart, expanding the portfolio of foundation models available to AWS customers. This model from Mistra

<p>Amazon SageMaker HyperPod now supports custom AMIs for Slurm-orchestrated clusters, enabling customers to deploy clusters with pre-configured, security-hardened environments that meet their specifi

<p>Amazon EMR on EKS now supports the Apache Spark troubleshooting agent. Data engineers can now diagnose EMR on EKS job failures through natural language, receiving automated root cause analysis and

<p>Amazon SageMaker HyperPod now supports AMI-based configuration for Slurm clusters that use continuous provisioning. Continuous provisioning adds nodes to the cluster as capacity becomes available,

<p>Amazon SageMaker Unified Studio now supports custom asset types for IAM-based domains. With custom asset types, domain administrators can catalog any format of asset within the SageMaker Unified St

<p>Amazon SageMaker HyperPod now supports deep health checks for Slurm-orchestrated clusters created with continuous provisioning, enabling you to proactively verify GPU accelerator health on running

<p>Amazon SageMaker Feature Store is a fully managed capability that makes it easy to compute, store, and retrieve features for training and deploying AI models. SageMaker Feature Store now supports n

<p>Amazon SageMaker Unified Studio now supports OpenLineage compatible data lineage in IAM-based domains, capturing lineage events from Apache Spark jobs run on Amazon EMR, AWS Glue, SageMaker Visual

<p>Amazon SageMaker Unified Studio now supports connecting existing Amazon Managed Workflows for Apache Airflow (MWAA) environments to projects. Data engineers and platform teams who already operate M

<p>Amazon SageMaker HyperPod now supports Disaggregated Prefill and Decode (DPD), an inference optimization that separates the two phases of large language model (LLM) inference — prefill and decode —

<p>Amazon SageMaker Unified Studio now supports Terraform for provisioning. Customers can use the open-source <a href="https://github.com/aws-ia/terraform-aws-sagemaker-unified-studio">terraform-aws-s

<p>Amazon SageMaker HyperPod now gives you visibility into the Amazon Machine Image (AMI) versions running across your clusters and automatically applies security patches without disrupting your workl

<p>Amazon SageMaker Inference now supports container image caching, enabling up to 2x faster end-to-end scaling for generative AI models during scale-out events. When your endpoint scales out, the ser

<p>Amazon SageMaker AI now supports serverless model customization for Gemma 4 E4B and 31B models using supervised fine-tuning (SFT), direct preference optimization (DPO), and reinforcement fine-tunin

<p>Amazon Managed Workflows for Apache Airflow (Amazon MWAA) Serverless now supports shared VPC subnets. Previously, customers using subnets shared via AWS Resource Access Manager (AWS RAM) received a

Amazon EC2 G6e インスタンスが SageMaker Notebook Instances で一般提供開始されました。G6e インスタンスは最大 8 個の NVIDIA L40s Tensor Core GPU（GPU あたり 48 GB メモリ）を搭載し、G5 インスタンスと比べて最大 2.5 倍のパフォーマンス向上を実現します。ユーザーはモデルのインタラクティブなテストやファインチューニング、13B パラメータまでの大規模言語モデルやディフュージョンモデルのデプロイに利用できます。

SageMaker Studio notebooksでAmazon EC2のG7eインスタンスタイプが利用可能になりました。G7eインスタンスは最大8個のNVIDIA RTX PRO 6000 Blackwell Server Edition GPU（GPU あたり96GB メモリ）と5世代Intel Xeonプロセッサを搭載しており、最大192vCPUと最大1600 Gbpsのネットワーク帯域幅に対応しています。米国東部（バージニア州・オハイオ州）および米国西部（オレゴン州）のリージョンで利用可能です。

AWSはall-MiniLM-L12-v2モデルをAmazon SageMaker JumpStartで提供開始しました。このモデルは文や段落を384次元の密ベクトル空間にマッピングし、セマンティック検索、テキストクラスタリング、文類似度判定アプリケーションの構築を可能にします。コンパクトな設計により高速推論を実現しながら、埋め込み品質を維持しています。

Mistral AI社の14Bパラメータのマルチモーダル基盤モデル「Ministral-3-14B-Instruct-2512」がAmazon SageMaker JumpStartで利用可能になりました。このモデルは画像解析、ネイティブ関数呼び出しによるエージェント機能、英語・日本語・中国語を含む多言語対応を備えています。SageMaker JumpStartのクリック数回の操作でデプロイでき、エッジ展開に最適化された軽量設計となっています。

<p>Amazon SageMaker AI's new observability capability allows customers to operate production generative AI inference workloads with confidence by providing comprehensive visibility into token performa

<p>AWS Glue Interactive Sessions now support Apache Spark Connect, using which you can now develop and run Apache Spark applications from your preferred environment, including managed notebooks in Ama