dosemeion
Results
1
comments of
dosemeion
不能使用batchsize=1，yolov5里面的yolo.py改成m.stride = torch.tensor([s / x.shape[-2] for x in self.forward(torch.zeros(2, ch, s, s))]) # forward