Yolov5: Label error while training with 2017 coco dataset. labels missing

[AlbertMP]

I downloaded the coco2017labels.zip from google drive and then unzipped it. When I tried to train the model with the entire coco dateset, the code reports error below that some images are missing in both train and val folder.

Optimizer groups: 62 .bias, 70 conv.weight, 59 other
Scanning labels ../coco/labels/train2017.cache (117266 found, 0 missing, 1021 empty, 0 duplicate, for 118287 images): 100%|█████████████████████████████████████| 118287/118287 [00:09<00:00, 13110.03it/s]
Scanning labels ../coco/labels/val2017.cache (4952 found, 0 missing, 48 empty, 0 duplicate, for 5000 images): 100%|█████████████████████████████████████████████████| 5000/5000 [00:00<00:00, 13120.22it/s]

I checked the ../coco/labels/train2017 and ../coco/labels/val2017 and found they just contain 117266 and 4952 labels.

Is the label files missing?

[github-actions[bot]]

Hello @AlbertMP, thank you for your interest in our work! Please visit our Custom Training Tutorial to get started, and see our Jupyter Notebook , Docker Image, and Google Cloud Quickstart Guide for example environments.

If this is a bug report, please provide screenshots and minimum viable code to reproduce your issue, otherwise we can not help you.

If this is a custom model or data training question, please note Ultralytics does not provide free personal support. As a leader in vision ML and AI, we do offer professional consulting, from simple expert advice up to delivery of fully customized, end-to-end production solutions for our clients, such as:

• Cloud-based AI systems operating on hundreds of HD video streams in realtime.
• Edge AI integrated into custom iOS and Android apps for realtime 30 FPS video inference.
• Custom data training, hyperparameter evolution, and model exportation to any destination.

For more information please visit https://www.ultralytics.com.

[NanoCode012]

My coco dataset reports the same number of label files. Could you check that you downloaded the entirety of the dataset? Around 20 GB. Did you download using the get_coco2017.sh file?

[AlbertMP]

My coco dataset reports the same number of label files. Could you check that you downloaded the entirety of the dataset? Around 20 GB. Did you download using the get_coco2017.sh file?

Yes. My train2017 folder contains 118287 pictures and val2017 folder contains 5000 pictures. It's the same as the number of dataset.
But some label files are missing I found when I unzipped the coco2017labels.zip file. That's why causing the problem I think.

You can check the number of label files, is yours equal to the number of pictures in train and val folders?

[NanoCode012]

Train labels : 117266
Valid labels : 4952
Train pic : 118287
Valid pic : 5000

However, I can run the code without any problem. I am not sure if the lack of labels just means that there is nothing to be detected in a picture.

Edit : Do you have problem running the code?

[glenn-jocher]

@NanoCode012 @AlbertMP not all images require labels. COCO has over 1000 unlabelled images.