AWS Machine Learning Blog

SEPTEMBER 18, 2024

The compute clusters used in these scenarios are composed of more than thousands of AI accelerators such as GPUs or AWS Trainium and AWS Inferentia , custom machine learning (ML) chips designed by Amazon Web Services (AWS) to accelerate deep learning workloads in the cloud.

Clustering 119

Clustering AWS ML ML 119

AWS Machine Learning Blog

MAY 1, 2024

Llama2 by Meta is an example of an LLM offered by AWS. To learn more about Llama 2 on AWS, refer to Llama 2 foundation models from Meta are now available in Amazon SageMaker JumpStart. Virginia) and US West (Oregon) AWS Regions, and most recently announced general availability in the US East (Ohio) Region.

AWS 127

AWS ML ML Clustering 127

AWS Machine Learning Blog

MAY 31, 2023

With containers, scaling on a cluster becomes much easier. In late 2022, AWS announced the general availability of Amazon EC2 Trn1 instances powered by AWS Trainium accelerators, which are purpose built for high-performance deep learning training. Therefore, we have two different options. Amazon Linux 2) ????????'

AWS 117

AWS Machine Learning Machine Learning ML 117

Flipboard

APRIL 13, 2023

Prime Air (our drones) and the computer vision technology in Amazon Go (our physical retail experience that lets consumers select items off a shelf and leave the store without having to formally check out) use deep learning. AWS has the broadest and deepest portfolio of AI and ML services at all three layers of the stack.

AWS 182

AWS ML ML AI 182

AWS Machine Learning Blog

OCTOBER 5, 2023

In this post, we walk through how to fine-tune Llama 2 on AWS Trainium , a purpose-built accelerator for LLM training, to reduce training times and costs. We review the fine-tuning scripts provided by the AWS Neuron SDK (using NeMo Megatron-LM), the various configurations we used, and the throughput results we saw.

AWS 128

AWS Machine Learning Machine Learning Deep Learning 128

AWS Machine Learning Blog

APRIL 29, 2024

For AWS and Outerbounds customers, the goal is to build a differentiated machine learning and artificial intelligence (ML/AI) system and reliably improve it over time. First, the AWS Trainium accelerator provides a high-performance, cost-effective, and readily available solution for training and fine-tuning large models.

AWS 127

AWS ML ML Python 127

Smart Data Collective

FEBRUARY 20, 2022

AWS (Amazon Web Services), the comprehensive and evolving cloud computing platform provided by Amazon, is comprised of infrastructure as a service (IaaS), platform as a service (PaaS) and packaged software as a service (SaaS). With its wide array of tools and convenience, AWS has already become a popular choice for many SaaS companies.

AWS 138

AWS Cloud Computing Data Lakes Database 138

AWS Machine Learning Blog

FEBRUARY 1, 2023

Recent developments in deep learning have led to increasingly large models such as GPT-3, BLOOM, and OPT, some of which are already in excess of 100 billion parameters. Many enterprise customers choose to deploy their deep learning workloads using Kubernetes—the de facto standard for container orchestration in the cloud.

AWS 98

AWS Clustering Deep Learning Deep Learning 98

AWS Machine Learning Blog

DECEMBER 12, 2023

In this post, we’ll summarize training procedure of GPT NeoX on AWS Trainium , a purpose-built machine learning (ML) accelerator optimized for deep learning training. M tokens/$) trained such models with AWS Trainium without losing any model quality. We’ll outline how we cost-effectively (3.2 billion in Pythia.

AWS 127

AWS Machine Learning Machine Learning Deep Learning 127

AWS Machine Learning Blog

JUNE 25, 2024

Amazon Web Services is excited to announce the launch of the AWS Neuron Monitor container , an innovative tool designed to enhance the monitoring capabilities of AWS Inferentia and AWS Trainium chips on Amazon Elastic Kubernetes Service (Amazon EKS). The Container Insights dashboard also shows cluster status and alarms.

AWS 97

AWS ML ML Clustering 97

AWS Machine Learning Blog

JUNE 27, 2024

In this post, we describe how we built our cutting-edge productivity agent NinjaLLM, the backbone of MyNinja.ai, using AWS Trainium chips. For training, we chose to use a cluster of trn1.32xlarge instances to take advantage of Trainium chips. We used a cluster of 32 instances in order to efficiently parallelize the training.

AWS 130

AWS AI AI Clustering 130

Flipboard

FEBRUARY 16, 2023

Modern model pre-training often calls for larger cluster deployment to reduce time and cost. In October 2022, we launched Amazon EC2 Trn1 Instances , powered by AWS Trainium , which is the second generation machine learning accelerator designed by AWS. The following diagram shows an example.

Clustering 132

Clustering AWS Deep Learning Deep Learning 132

AWS Machine Learning Blog

JUNE 11, 2024

In this post, we describe the scale of our AI offerings, the challenges with diverse AI workloads, and how we optimized mixed AI workload inference performance with AWS Graviton3 based c7g instances and achieved 20% throughput improvement, 30% latency reduction, and reduced our cost by 25–30%.

Machine Learning 122

Machine Learning Machine Learning AWS Natural Language Processing 122

AWS Machine Learning Blog

JANUARY 17, 2024

Today, we’re excited to announce the availability of Llama 2 inference and fine-tuning support on AWS Trainium and AWS Inferentia instances in Amazon SageMaker JumpStart. In this post, we demonstrate how to deploy and fine-tune Llama 2 on Trainium and AWS Inferentia instances in SageMaker JumpStart.

AWS 131

AWS Python Machine Learning Machine Learning 131

AWS Machine Learning Blog

SEPTEMBER 18, 2024

Zeta’s AI innovations over the past few years span 30 pending and issued patents, primarily related to the application of deep learning and generative AI to marketing technology. Zeta’s AI innovation is powered by a proprietary machine learning operations (MLOps) system, developed in-house.

AWS 123

AWS Machine Learning Machine Learning ML 123

AWS Machine Learning Blog

NOVEMBER 30, 2023

The number of companies launching generative AI applications on AWS is substantial and building quickly, including adidas, Booking.com, Bridgewater Associates, Clariant, Cox Automotive, GoDaddy, and LexisNexis Legal & Professional, to name just a few. Innovative startups like Perplexity AI are going all in on AWS for generative AI.

AWS 145

AWS AI AI ML 145

AWS Machine Learning Blog

MAY 31, 2024

In this blog post and open source project , we show you how you can pre-train a genomics language model, HyenaDNA , using your genomic data in the AWS Cloud. Amazon SageMaker Amazon SageMaker is a fully managed ML service offered by AWS, designed to reduce the time and cost associated with training and tuning ML models at scale.

AWS 119

AWS ML ML Machine Learning 119

AWS Machine Learning Blog

APRIL 1, 2024

Machine learning (ML) research has proven that large language models (LLMs) trained with significantly large datasets result in better model quality. Distributed model training requires a cluster of worker nodes that can scale. The following figure shows how FSDP works for two data parallel processes.

Clustering 131

Clustering AWS ML ML 131

AWS Machine Learning Blog

JUNE 7, 2023

Libraries such as DeepSpeed (an open-source deep learning optimization library for PyTorch) address some of these challenges, and can help accelerate model development and training. Training setup We provisioned a managed compute cluster comprised of 16 dl1.24xlarge instances using AWS Batch. Pre-training of a 1.5-billion-parameter

AWS 97

AWS Clustering Deep Learning Deep Learning 97

AWS Machine Learning Blog

SEPTEMBER 7, 2023

Since then, this feature has been integrated into many of our managed Amazon Machine Images (AMIs), such as the Deep Learning AMI and the AWS ParallelCluster AMI. Prerequisites To simplify reproducing the entire stack from this post, we use a container that has all the required tooling (aws cli, eksctl, helm, etc.)

Clustering 126

Clustering AWS Machine Learning Machine Learning 126

AWS Machine Learning Blog

NOVEMBER 22, 2023

Similar to the rest of the industry, the advancements of accelerated hardware have allowed Amazon teams to pursue model architectures using neural networks and deep learning (DL). Last year, AWS launched its AWS Trainium accelerators, which optimize performance per cost for developing and building next generation DL models.

AWS 115

AWS ML ML Deep Learning 115

AWS Machine Learning Blog

APRIL 19, 2024

Our deep learning models have non-trivial requirements: they are gigabytes in size, are numerous and heterogeneous, and require GPUs for fast inference and fine-tuning. The architecture deploys a simple service in a Kubernetes pod within an EKS cluster. The following diagram illustrates the solution architecture.

Clustering 127

Clustering AI AI AWS 127

AWS Machine Learning Blog

DECEMBER 22, 2023

As a result, machine learning practitioners must spend weeks of preparation to scale their LLM workloads to large clusters of GPUs. Integrating tensor parallelism to enable training on massive clusters This release of SMP also expands PyTorch FSDP’s capabilities to include tensor parallelism techniques.

Clustering 123

Clustering Deep Learning Deep Learning AWS 123

AWS Machine Learning Blog

MAY 23, 2024

By distributing experts across workers, expert parallelism addresses the high memory requirements of loading all experts on a single device and enables MoE training on a larger cluster. The following figure offers a simplified look at how expert parallelism works on a multi-GPU cluster.

Clustering 132

Clustering AWS Deep Learning Deep Learning 132

AWS Machine Learning Blog

JANUARY 13, 2023

To mitigate these challenges, we propose a federated learning (FL) framework, based on open-source FedML on AWS, which enables analyzing sensitive HCLS data. It involves training a global machine learning (ML) model from distributed health data held locally at different sites. Request a VPC peering connection.

AWS 101

AWS Analytics Analytics Machine Learning 101

AWS Machine Learning Blog

OCTOBER 23, 2024

SnapLogic uses Amazon Bedrock to build its platform, capitalizing on the proximity to data already stored in Amazon Web Services (AWS). To address customers’ requirements about data privacy and sovereignty, SnapLogic deploys the data plane within the customer’s VPC on AWS.

AI 92

AI AI Database AWS 92

AWS Machine Learning Blog

MARCH 11, 2025

The MoE architecture allows activation of 37 billion parameters, enabling efficient inference by routing queries to the most relevant expert clusters. By integrating this model with Amazon SageMaker AI , you can benefit from the AWS scalable infrastructure while maintaining high-quality language model capabilities.

AWS 83

AWS ML ML Natural Language Processing 83

Flipboard

MARCH 9, 2023

In this post, we review the technical requirements and application design considerations for fine-tuning and serving hyper-personalized AI models at scale on AWS. Second, SageMaker supports unique GPU-enabled hosting options for deploying deep learning models at scale.

AWS 149

AWS AI AI ML 149

AWS Machine Learning Blog

SEPTEMBER 21, 2023

To add to our guidance for optimizing deep learning workloads for sustainability on AWS , this post provides recommendations that are specific to generative AI workloads. Adopt an efficient inference infrastructure – You can deploy your models on an AWS Inferentia2 accelerator.

AWS 134

AWS AI AI Deep Learning 134

AWS Machine Learning Blog

SEPTEMBER 11, 2024

AWS published Guidance for Optimizing MLOps for Sustainability on AWS to help customers maximize utilization and minimize waste in their ML workloads. In this blog post, you will learn how to optimize MLOps for sustainability. Start with the AWS Region you choose for your workload.

AWS 119

AWS Data Preparation ML ML 119

AWS Machine Learning Blog

AUGUST 8, 2024

Webex’s focus on delivering inclusive collaboration experiences fuels their innovation, which uses artificial intelligence (AI) and machine learning (ML), to remove the barriers of geography, language, personality, and familiarity with technology. Webex works with the world’s leading business and productivity apps—including AWS.

AWS 127

AWS AI AI Clustering 127

AWS Machine Learning Blog

MARCH 2, 2023

Hyperparameter optimization is highly computationally demanding for deep learning models. In our solution, we implement a hyperparameter grid search on an EKS cluster for tuning a bert-base-cased model for classifying positive or negative sentiment for stock market data headlines. The code can be found on the GitHub repo.

Clustering 98

Clustering AWS Deep Learning Deep Learning 98

AWS Machine Learning Blog

MARCH 6, 2023

These factors require training an LLM over large clusters of accelerated machine learning (ML) instances. In the past few years, numerous customers have been using the AWS Cloud for LLM training. We recommend working with your AWS account team or contacting AWS Sales to determine the appropriate Region for your LLM workload.

AWS 107

AWS Clustering ML ML 107

AWS Machine Learning Blog

APRIL 19, 2024

We are excited to announce a new version of the Amazon SageMaker Operators for Kubernetes using the AWS Controllers for Kubernetes (ACK). ACK is a framework for building Kubernetes custom controllers, where each controller communicates with an AWS service API. They are also supported by AWS CloudFormation. Release v1.2.9

AWS 130

AWS ML ML Machine Learning 130

AWS Machine Learning Blog

OCTOBER 2, 2024

In this post, we share AWS guidance that we have learned and developed as part of real-world projects into practical guides oriented towards the AWS Well-Architected Framework , which is used to build production infrastructure and applications on AWS. We focus on the operational excellence pillar in this post.

AWS 131

AWS AI AI ML 131

AWS Machine Learning Blog

DECEMBER 5, 2023

Walkthrough AWS-optimized AllGather AWS-optimized AllGather uses the following techniques to achieve better performance on AWS infrastructure compared to NCCL: We move data between instances via Elastic Fabric Adapter (EFA) network with an all-to-all communication pattern. 24xlarge nodes (512 NVIDIA A100 GPUs) PyTorch FSDP 97.89

AWS 124

AWS Deep Learning Deep Learning Clustering 124

AWS Machine Learning Blog

SEPTEMBER 12, 2024

In this post, we explore the journey that Thomson Reuters took to enable cutting-edge research in training domain-adapted large language models (LLMs) using Amazon SageMaker HyperPod , an Amazon Web Services (AWS) feature focused on providing purpose-built infrastructure for distributed training at scale. So, for example, a 6.6B

Clustering 115

Clustering AWS ML ML 115