1

I am trying to build an object detection system using YOLOv5x. I created my dataset via the Roboflow website and after downloading the set to my notebook, I am running this command to start training:

!curl -L "https://app.roboflow.com/ds/[DATASET-LINK]" > roboflow.zip; unzip roboflow.zip; rm roboflow.zip

But I am getting this error for every single image:

train: WARNING: Ignoring corrupted image and/or label /content/yolov5/train/images/output_0446_png.rf.caced7dfbd9c68fe51180ceb8c2f04e8.jpg: duplicate labels
train: WARNING: Ignoring corrupted image and/or label /content/yolov5/train/images/output_0450_png.rf.808e3c83dd6b516900257848467d9a5b.jpg: duplicate labels
train: WARNING: Ignoring corrupted image and/or label /content/yolov5/train/images/output_0456_png.rf.898ad055d9c4cf67db7657c4901db2b7.jpg: duplicate labels
train: WARNING: Ignoring corrupted image and/or label /content/yolov5/train/images/output_0459_png.rf.8bc9567fac8542598a79c2bf11d4d8d5.jpg: duplicate labels
train: WARNING: Ignoring corrupted image and/or label /content/yolov5/train/images/output_0461_png.rf.62c902b73e1b6a92e1417b90c8dd3c9c.jpg: duplicate labels
train: WARNING: Ignoring corrupted image and/or label /content/yolov5/train/images/output_0462_png.rf.bf025028cd9eb5fe98d3cd80452a8d86.jpg: duplicate labels
train: WARNING: Ignoring corrupted image and/or label /content/yolov5/train/images/output_0463_png.rf.931cea20da666391b1b855eed678dc47.jpg: duplicate labels
train: WARNING: Ignoring corrupted image and/or label /content/yolov5/train/images/output_0465_png.rf.5bedf83f99b1ee0bf8116c2d369ba202.jpg: duplicate labels
train: WARNING: Ignoring corrupted image and/or label /content/yolov5/train/images/output_0466_png.rf.75ceb50e893e1d7ff43a49c40767d552.jpg: duplicate labels

Its complaining about "duplicate labels" but I have no idea what it means by this. I do not have 2 of the same classes and this dataset worked well with YOLOv4 and YOLOv4-Tiny.

4

1 回答 1

4

I also encountered this problem when I used yolov5. It's mostly during data preprocessing. And the reason I found out was that, as the tip says, the label was duplicated. In the txt tag file, the location of the label is repeated. As shown below, the tags are duplicated four times. If only once, the problem disappears. enter image description here

于 2021-03-16T01:21:04.873 回答