Hi, How can I solve this problem

[pureb1ack1]

File "C:/Users/Administrator/Desktop/CSGNet-master/train_synthetic.py", line 80, in
imitate_net.load_state_dict(torch.load(config.pretrain_modelpath))
File "D:\anaconda3\envs\cv\lib\site-packages\torch\nn\modules\module.py", line 490, in load_state_dict
.format(name))
KeyError: 'unexpected key "encoder.conv4.weight" in state_dict'

[pureb1ack1]

This happened when I used pretrain_model, can I not use pretrain_ model for training?