supervision: Loading OBB detection from ultralytics fails

Search before asking

  • I have searched the Supervision issues and found no similar bug report.

Bug

When passing in the results from ultralytics yolov8 OBB model to from_ultralytics it will fail and give the following:

AttributeError: ‘NoneType’ object has no attribute ‘cls’

due to:

class_id = ultralytics_results.boxes.cls.cpu().numpy().astype(int)

Environment

  • Supervision 0.20.0
  • Ultralytics 8.2.18

Minimal Reproducible Example

from typing import List

import cv2
from ultralytics import YOLO
from supervision import Detections
from ultralytics.engine.results import Results
from supervision.tracker.byte_tracker.core import ByteTrack

yolo = YOLO("yolov8n-obb.pt", task="obb")

tracker = ByteTrack(track_activation_threshold=0.8,
                    minimum_matching_threshold=0.8,
                    lost_track_buffer=30,
                    frame_rate=30)

video_feed = cv2.VideoCapture("test_rgb.mp4")

status, frame = video_feed.read()

while status:
    results: List[Results] = yolo(frame,
                                  show=False,
                                  verbose=False,
                                  imgsz=[736, 1280])
    result: Results = results[0]

    inference_detections: Detections = Detections.from_ultralytics(result)
    detections = tracker.update_with_detections(inference_detections)

    for (xyxy, mask, conf, class_id, tracker_id, data) in detections:
        print("xyxy:", xyxy)

    status, frame = video_feed.read()

Additional

No response

Are you willing to submit a PR?

  • Yes I’d like to help by submitting a PR!

About this issue

  • Original URL
  • State: closed
  • Created 3 years ago
  • Comments: 18 (4 by maintainers)

Most upvoted comments