mickey-cool

I encountered this problem when I try to train on a single Gpu, please help me,Thank you : python main_persformer.py --mod=kkk --batch_size=8 --nepochs=200 proc_id: None world size: 1 local_rank: None...

The result is different before and after reparametrization

7

After reparametrization of yolov7-tiny, more bboxes came out, and confidence boosted, any idea? Before：  After:  **my reparemeterization code:** `# import from copy import deepcopy from models.yolo import Model...