Description
The Product: AWS Machine Learning accelerators are at the forefront of AWS innovation. The Inferentia chip delivers best-in-class ML inference performance at the lowest cost in cloud. Trainium will deliver the best-in-class ML training performance with the most teraflops (TFLOPS) of compute power for ML in the cloud. This is all enabled by edge software stack, the AWS Neuron Software Development Kit (SDK), which includes an ML compiler, Neuron Kernel Interface (NKI) compiler, and runtime that natively integrates into popular ML frameworks, such as PyTorch and TensorFlow.

Neuron Kernel Interface (NKI) is a bare-metal language and compiler for directly programming NeuronDevices available on AWS Trn/Inf instances. You can use NKI to develop, optimize and run new operators directly on NeuronCores while making full use of available compute and memory resources.

AWS Neuron and Inferentia are used at scale with customers and partners like PyTorch, Epic Games, ...