keras-cv icon indicating copy to clipboard operation
keras-cv copied to clipboard

Add YOLOv8-Pose

Open josht000 opened this issue 1 year ago • 3 comments

Short Description

Add support for the various size YOLOv8 pose estimation models.

  • [ ] inference scripts for pretrained models if any
  • [ ] training scripts/examples
  • [ ] support custom key-point configurations. Ie. not just keypoints for people but also cars, etc.

Existing Implementations

  • https://github.com/jahongir7174/YOLOv8-pose
  • https://github.com/ultralytics/ultralytics/blob/main/ultralytics/nn/modules/head.py#L110

josht000 avatar Aug 24 '23 13:08 josht000

@josht000 thank you for the suggestion. This is not on our priority list for the team at the moment. I will leave the issue open. However, if you do have a PR ready for review please do submit it. We will review it.

divyashreepathihalli avatar Feb 05 '24 19:02 divyashreepathihalli

@divyashreepathihalli thanks for the feedback, are there plans to support any keypoint models in keras cv?

barrypitman avatar Feb 05 '24 20:02 barrypitman

@barrypitman check out or 3d-OD model - https://github.com/keras-team/keras-cv/tree/1def8ba37763fa5185761bd0a266fb9acda590c2/keras_cv/models/object_detection_3d

divyashreepathihalli avatar Feb 12 '24 18:02 divyashreepathihalli