github-day
commented
1 year ago Thank you for your work. I have some questions about your project code: I looked at the model you called YoloV5 in the middle, but did not see the operation about model modification mentioned in your article (modifying the image network structure, such as modifying the Stride size, upsampling the image, and Ensemble). The most important thing is that you do not read the model. yaml configuration file like YOLOv5, how did you switch the detection framework of YOLOV5 to YOLT?
Have you figured it out? Please let me know.
I read train.py and yolo.py, finding that it's just using the default yolov5l.pt and yolov5l.yaml by ckpt = torch.load(weights, map_location='cpu') #train.py line 119.
Maybe the original yolov5l model is good enough???
Thank you for your work. I have some questions about your project code: I looked at the model you called YoloV5 in the middle, but did not see the operation about model modification mentioned in your article (modifying the image network structure, such as modifying the Stride size, upsampling the image, and Ensemble). The most important thing is that you do not read the model. yaml configuration file like YOLOv5, how did you switch the detection framework of YOLOV5 to YOLT?