Great to hear it’s working!
For anyone landing here later: when using OpenVINO exports, load with the correct task (detect/segment/classify/pose/obb) and either match the training resolution at inference with imgsz or export with dynamic=True. The workflow and options are summarized in the Ultralytics OpenVINO export guide.
If anything else pops up, please share yolo checks and the exact error so we can help quickly.