Skip to content

Support Deployment for serving single-node models #32

Open
@kerthcet

Description

@kerthcet

We support lws as the default workload, however, most of the cases mutli-hosts is not needed, even with Llama3.1 405B. So maybe this is a better choice, people don't need to lws controller then. However, this leads to the complexities of workload orchestrations.

Metadata

Metadata

Assignees

No one assigned

    Labels

    backlogHigher priority than priority/awaiting-more-evidence.needs-triageIndicates an issue or PR lacks a label and requires one.questionFurther information is requested

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions