The AMD ROCm blog just announced a new open-source AMD AI software project: the AMD Enterprise AI Suite as well as AMD Inference Microservices (AIMs).
With the flurry of AMD AI-related software announcements over the past number of months and the growing work around ROCm, you may be wondering — like I — what is this “AMD Enterprise AI Suite” now… AMD Enterprise AI Suite is complementary to ROCm and builds off it as the base. This open-source AMD Enterprise AI Suite is described as being a “full stack solution” for developing, deploying, and running AI workloads on AMD Instinct hardware. The AMD Enterprise AI Suite is Kubernetes focused and is intended for use by Linux server administrators to AI researchers and software developers.
The AMD Enterprise AI Suite aims to optimize GPU efficiency, unify AI infrastructure, accelerate time-to-production, and provide AI-native workload orchestration around Kubernetes. It’s very Kubernetes focused from my initial reading of the new documentation on the AMD Enterprise AI Suite.
As part of the AMD Enterprise AI Suite is the AMD AI Workbench for providing various AMD Inference Microservices (AIMs) that provide for easy development and deployment of different models and other AI software components. AMD Enterprise AI Suite appears to be built off Cluster-Forge that was developed by Silo AI as one of the AI companies that AMD acquired last year.
More details on the AMD Enterprise AI Suite can be found via the ROCm blog, AMD announcement, and Enterprise AI Suite documentation site.
