search

WongKinYiu / yolov9

Implementation of paper - YOLOv9: Learning What You Want to Learn Using Programmable Gradient Information
GNU General Public License v3.0
9.02k stars 1.43k forks source link

Hello, can you provide the complete code of "YOLOR-BASED MULTI-TASK LEARNING"? #467

Open coder678 opened 5 months ago

WongKinYiu commented 5 months ago

The code are already released except for image captioning.

coder678 commented 5 months ago

Thank you for your reply and great work. I tried running the model and found the following problems in the semantic segmentation part: (1) I ran python /yolov9/panoptic/predict.py --weight gelan-c-pan.pt --data /yolov9/data/coco.yaml the model only outputs the content of target detection and instance segmentation, and lacks semantic segmentation, which is the same as /yolov9/segment/predict.py outputs the same result (2) I ran python /yolov9/panoptic/val.py --weight gelan-c-pan.pt --data /yolov9/data/coco.yaml the model correctly outputs the label images of target detection, instance segmentation, and semantic segmentation, but the predict semantic segmentation content is a pink solid color picture. Through debugging, I found that the predicted semantic segmentation (plot_semasks) is very different from the label semantic segmentation (semasks).The problem is as shown below image image

WongKinYiu commented 5 months ago

Oh, we not yet update the visualization code. We will update them soon.

coder678 commented 5 months ago

OK, thanks