-
Notifications
You must be signed in to change notification settings - Fork 175
Open
Description
为什么在20个类别上的数据集上进行微调过后进行指定类别推理的时候就报错了,中间特征图的大小被固定成训练的类别了这是什么原因牙官方预训练过后的就不会有这个问题。x_cat = torch.cat([xi.view(shape[0], self.nc + self.reg_max * 4, -1) for xi in x], 2)
File "/home/xgrids/ZJJ/code/yoloe/ultralytics/nn/modules/head.py", line 486, in
x_cat = torch.cat([xi.view(shape[0], self.nc + self.reg_max * 4, -1) for xi in x], 2)
RuntimeError: shape '[1, 65, -1]' is invalid for input of size 921600
Metadata
Metadata
Assignees
Labels
No labels