A high-throughput and memory-efficient inference and serving engine for LLMs

Large-scale LLM inference engine

Foundation model benchmarking tool. Run any model on any AWS platform and benchmark for performance across instance type and serving stack options.

This Guidance demonstrates how to deploy a machine learning inference architecture on Amazon Elastic Kubernetes Service (Amazon EKS). It addresses the basic implementation requirements as well as ways you can pack thousands of unique PyTorch deep learning (DL) models into a scalable architecture and evaluate performance

CMP314 Optimizing NLP models with Amazon EC2 Inf1 instances in Amazon Sagemaker

Collection of bet practices, reference architectures, examples, and utilities for foundation model development and deployment on AWS.

This repository provides an easy hands-on way to get started with AWS Inferentia. A demonstration of this hands-on can be seen in the AWS Innovate 2023 - AIML Edition session.

Sentence Transformers on EC2 Inf1

Deploy Large Models on AWS Inferentia (Inf2) instances.