About HAMi
HAMi helps Kubernetes clusters share and schedule heterogeneous AI computing devices.
What is HAMi?
HAMi (Heterogeneous AI Computing Virtualization Middleware) is an open-source project that provides device sharing, fine-grained resource allocation, and device-aware scheduling for GPUs, NPUs, and other accelerators in Kubernetes. It was formerly known as k8s-vGPU-scheduler.
HAMi is a CNCF Incubating project. For technical capabilities and usage, start with the HAMi documentation.
Open governance
HAMi is developed by an open community. Project roles, decision-making, and maintainer responsibilities are documented in the governance guide, and community participation follows the CNCF Code of Conduct. Visit the community page to join discussions, meetings, and contributions.
Contribute
HAMi welcomes community contributions. Start with the contributor guide to report issues, propose changes, and follow the project's GitHub workflow.
Licenses
HAMi code and this website's content use different licenses. The HAMi software is licensed under Apache License 2.0; this website and its documentation are licensed under Creative Commons Attribution 4.0 International (CC BY 4.0).
