Burhan

@KozlovKY I can't recall the device or versions I used, but in our Discord server, I shared with a user my test results. Once you join the server and select...

@lakshanthad maybe this is something you'd like to take a crack at?

Thanks for letting us know @johnnynunez 🚀

@elisno Thank you for the reply and example(s). > It's unclear what the -1 label represents in your example This was a placeholder and I didn't mean to include it,...

If my understanding of your explanation is correct, the array would likely need to be constructed using something like: ```py K = 368 # Number of classes # example detection...

Well, my conclusion was incorrect and `list[list[np.ndarray]]` doesn't work. ``` Traceback (most recent call last): File "", line 1, in is_issue = _find_label_issues( ^^^^^^^^^^^^^^^^^^^ File ".venv/Lib/site-packages/cleanlab/object_detection/filter.py", line 138, in _find_label_issues...

Very curious! Benchmarks are failing with this change. The `mAP50-95(M)` on all have dropped by $0.018$ but PyTorch has dropped $0.06$ for YOLOv8-seg ## On this PR Benchmarks (ubuntu-latest, 3.11,...

1 pixel difference in resizing, makes for quite the difference in the benchmark result. This result is not intuitive and it's not clear how 1 pixel would change the results...

Testing on COCO validation (detect) `yolo val model=yolov8n.pt data=coco.yaml` ## Branch result ``` Class Images Instances Box(P R mAP50 mAP50-95) all 5000 36335 0.633 0.475 0.521 0.371 person 2693 10777...

Converted to draft. Leaving to continue testing when time permits. Will share learnings/findings here for future reference.