剪枝后微调 加载模型的时候报错

[woshituobaye]

File "/home/nxin/下载/yolov5-v6-prune/prune_finetune.py", line 134, in train cfg_model.load_state_dict(torch.load(weights)['model'])

RuntimeError: Error(s) in loading state_dict for Darknet: size mismatch for module_list.72.Conv2d.weight: copying a param with shape torch.Size([18, 62, 1, 1]) from checkpoint, the shape in current model is torch.Size([255, 62, 1, 1]). size mismatch for module_list.72.Conv2d.bias: copying a param with shape torch.Size([18]) from checkpoint, the shape in current model is torch.Size([255]). size mismatch for module_list.84.Conv2d.weight: copying a param with shape torch.Size([18, 121, 1, 1]) from checkpoint, the shape in current model is torch.Size([255, 121, 1, 1]). size mismatch for module_list.84.Conv2d.bias: copying a param with shape torch.Size([18]) from checkpoint, the shape in current model is torch.Size([255]). size mismatch for module_list.96.Conv2d.weight: copying a param with shape torch.Size([18, 228, 1, 1]) from checkpoint, the shape in current model is torch.Size([255, 228, 1, 1]). size mismatch for module_list.96.Conv2d.bias: copying a param with shape torch.Size([18]) from checkpoint, the shape in current model is torch.Size([255]).

[ZJU-lishuang]

类别数检查一下

[gmt710]

@woshituobaye , Hello, have you solved this?

[gmt710]

知道了，需要自己修改下生成的cfg里面的filter