Installation#
Deployment options#
The AMD AI Workbench can be deployed in multiple configurations:
Standalone: AMD AI Workbench features operate independently with namespace-scoped access and basic authentication
Combined with AMD Resource Manager: Both AI Workbench and Resource Manager are deployed together, providing full integration with enterprise resource management, quotas, and organizational hierarchy for access to all features
The core AI development features (model deployment, fine-tuning, workspaces, chat) are available in all deployment modes. When integrated with AMD Resource Manager, additional enterprise features such as quota management, organizational hierarchy, and advanced access control become available.
Standalone mode#
In standalone mode, the AI Workbench is scoped to a single project namespace created during installation. All resources, including AIMs, workloads, workspaces, and datasets, are managed within this namespace.
Combined mode#
When deployed with AMD Resource Manager, the AI Workbench supports multiple project namespaces. Users can access projects they belong to, with AMD Resource Manager handling namespace lifecycle, quota allocation, and organizational hierarchy.
Install AMD AI Workbench#
The AMD AI Workbench can be installed using the following guides:
On-prem installation: Install both AMD AI Workbench and AMD Resource Manager
On-prem installation - Standalone: Install AI Workbench standalone
DigitalOcean installation: Install in the DigitalOcean cloud environment
Helm installation guide: This guide provides step-by-step instructions for deploying the AMD enterprise AI reference stack on a Kubernetes cluster using the Helm charts in the repository
After installation, continue with the Quick start and explore the UI.