Skip to content

Would centralized RK3588 infrastructure be useful for CV deployment/testing? #5

Description

@Josh-Difratech

We are evaluating whether a large centralized RK3588 fleet has a useful role in the CV ecosystem.
We control 1,280 identical RK3588 nodes already deployed in U.S. colocation.
This repository caught my attention because the Dockerized FastAPI architecture looks very close to something that could be run as a hosted service.

I’m trying to understand whether there is meaningful demand for any of the following:

  • hosted YOLO / classification / segmentation inference
  • concurrent video-stream processing
  • remote RK3588 hardware testing
  • model compatibility validation
  • performance benchmarking before deploying to edge devices

From your experience with customers/developers using RK3588, which of those problems actually exists today?

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