Skip to content

model config transformation should happen in Ray #23

@ben-directai

Description

@ben-directai

Summary:

See discussion in #21

We are storing a transformed dictionary in Redis to represent model configs. There's a lot of annoyance related to representing the configs correctly when we return them back to the client. Would be ideal if none of that transformation was happening in fastapi and we just passed a ClassifierDeploy or DetectorDeploy instance directly to Ray.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions