(28)改进的yolov4训练自制2400张数据集,预测报错

在使用改进的Yolov4训练2400张自制数据集后,遇到预测报错。通过排查确定不是数据集问题,而是训练模型的Attention模块与测试时的配置不符。训练模型采用ECA,但测试时误以为是SE,导致权重不匹配。尝试的解决方案包括修改加载权重的方式,但并未解决问题。最终,调整训练和测试中的Attention模块参数匹配,以提高预测精度。
摘要由CSDN通过智能技术生成

报错信息:

Traceback (most recent call last):
  File "/home/PycharmProjects/yolov4/3-L_m-lite-Att-later_marker2400/get_dr_txt.py", line 119, in <module>
    yolo = mAP_Yolo()
  File "/home/PycharmProjects/yolov4/3-L_m-lite-Att-later_marker2400/yolo.py", line 56, in __init__
    self.generate()
  File "/home/PycharmProjects/yolov4/3-L_m-lite-Att-later_marker2400/yolo.py", line 93, in generate
    self.net.load_state_dict(state_dict)
  File "/home/.conda/envs/ultralytics_yolov3/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1045, in load_state_dict
    self.__class__.__name__, "\n\t".join(error_msgs)))
RuntimeError: Error(s) in loading state_dict for YoloBody:
	Missing key(s) in state_dict: "feat0_att.fc.0.weight", "feat0_att.fc.2.weight", "feat1_att.fc.0.weight", "feat1_att.fc.2.weight", "feat2_att.fc.0.weight", "feat2_att.fc.2.weight", "upsample1_att.fc.0.weight", "upsample1_att.fc.2.weight", "upsample2_att.fc.0.weight", "upsample2_att.fc.2.weight". 
	Unexpected key(s) in state_dict: "feat0_att.convs.0.0.weight", "feat0_att.convs.0.0.bias", "feat0_att.convs.0.1.weight", "feat0_att.convs.0.1.bias", "feat0_att.convs.0.1.running_mean", "feat0_att.convs.0.1.running_var", "feat0_att.convs.0.1.num_batches_tracked", "feat0_att.convs.1.0.weight", "feat0_att.convs.1.0.bias", "feat0_att.convs.1.1.weight", "feat0_att.convs.1.1.bias", "fe
  • 1
    点赞
  • 0
    收藏
    觉得还不错? 一键收藏
  • 0
    评论
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值