I’m the co-founder of Onepanel.io open-core (Cloud-native OS for AI), and we are interested in supporting/developing functionality to enable annotation, GPU access via API, and computer vision pipeline support directly within the UI. We’ve worked with and provided open-source support for CVAT ( Video Annotation Tool) and would like to do the same with WebODM.
We’ve already created a YAML template with the corresponding Docker container allowing it to work with Onepanel CE workflow automation and Kubernetes-native infra.
The goal is to provide computer vision pre-annotation using a backend AI workflow to automatically train models for both object detection and semantic segmentation. Onepanel would provide the model training pipeline - we just need a working UI for annotation and inference in the UI of WebODM.
If you are interested in this please respond and help to define the requirements for the same. We will obviously release it openly.