测试报错

[foralliance]

@qijiezhao HI

测试时，执行
CUDA_VISIBLE_DEVICES=0 python test.py --config configs/cfenet300_vgg16.py -d COCO -m=weights/cfenet300_vgg16_coco.pth

报错：

Traceback (most recent call last):
  File "test.py", line 136, in <module>
    net.load_state_dict(new_state_dict)
  File "/home/jmx/.local/lib/python2.7/site-packages/torch/nn/modules/module.py", line 721, in load_state_dict
    self.__class__.__name__, "\n\t".join(error_msgs)))
RuntimeError: Error(s) in loading state_dict for CFENet:
	Unexpected key(s) in state_dict: "Norm.num_batches_tracked", "reduce1.bn.num_batches_tracked", "up_reduce1.bn.num_batches_tracked", "reduce2.bn.num_batches_tracked", "up_reduce2.bn.num_batches_tracked", "arterial.0.cfem_a.0.bn.num_batches_tracked", "arterial.0.cfem_a.1.bn.num_batches_tracked", "arterial.0.cfem_a.2.bn.num_batches_tracked", "arterial.0.cfem_a.3.bn.num_batches_tracked", "arterial.0.cfem_a.4.bn.num_batches_tracked", "arterial.0.cfem_b.0.bn.num_batches_tracked", "arterial.0.cfem_b.1.bn.num_batches_tracked", "arterial.0.cfem_b.2.bn.num_batches_tracked", "arterial.0.cfem_b.3.bn.num_batches_tracked", "arterial.0.cfem_b.4.bn.num_batches_tracked", "arterial.0.ConvLinear.bn.num_batches_tracked", "arterial.0.shortcut.bn.num_batches_tracked", "arterial.1.cfem_a.0.bn.num_batches_tracked", "arterial.1.cfem_a.1.bn.num_batches_tracked", "arterial.1.cfem_a.2.bn.num_batches_tracked", "arterial.1.cfem_a.3.bn.num_batches_tracked", "arterial.1.cfem_a.4.bn.num_batches_tracked", "arterial.1.cfem_b.0.bn.num_batches_tracked", "arterial.1.cfem_b.1.bn.num_batches_tracked", "arterial.1.cfem_b.2.bn.num_batches_tracked", "arterial.1.cfem_b.3.bn.num_batches_tracked", "arterial.1.cfem_b.4.bn.num_batches_tracked", "arterial.1.ConvLinear.bn.num_batches_tracked", "arterial.1.shortcut.bn.num_batches_tracked", "arterial.2.cfem_a.0.bn.num_batches_tracked", "arterial.2.cfem_a.1.bn.num_batches_tracked", "arterial.2.cfem_a.2.bn.num_batches_tracked", "arterial.2.cfem_a.3.bn.num_batches_tracked", "arterial.2.cfem_a.4.bn.num_batches_tracked", "arterial.2.cfem_b.0.bn.num_batches_tracked", "arterial.2.cfem_b.1.bn.num_batches_tracked", "arterial.2.cfem_b.2.bn.num_batches_tracked", "arterial.2.cfem_b.3.bn.num_batches_tracked", "arterial.2.cfem_b.4.bn.num_batches_tracked", "arterial.2.ConvLinear.bn.num_batches_tracked", "arterial.2.shortcut.bn.num_batches_tracked", "arterial.3.cfem_a.0.bn.num_batches_tracked", "arterial.3.cfem_a.1.bn.num_batches_tracked", "arterial.3.cfem_a.2.bn.num_batches_tracked", "arterial.3.cfem_a.3.bn.num_batches_tracked", "arterial.3.cfem_a.4.bn.num_batches_tracked", "arterial.3.cfem_b.0.bn.num_batches_tracked", "arterial.3.cfem_b.1.bn.num_batches_tracked", "arterial.3.cfem_b.2.bn.num_batches_tracked", "arterial.3.cfem_b.3.bn.num_batches_tracked", "arterial.3.cfem_b.4.bn.num_batches_tracked", "arterial.3.ConvLinear.bn.num_batches_tracked", "arterial.3.shortcut.bn.num_batches_tracked", "lateral.0.cfem_a.0.bn.num_batches_tracked", "lateral.0.cfem_a.1.bn.num_batches_tracked", "lateral.0.cfem_a.2.bn.num_batches_tracked", "lateral.0.cfem_a.3.bn.num_batches_tracked", "lateral.0.cfem_a.4.bn.num_batches_tracked", "lateral.0.cfem_b.0.bn.num_batches_tracked", "lateral.0.cfem_b.1.bn.num_batches_tracked", "lateral.0.cfem_b.2.bn.num_batches_tracked", "lateral.0.cfem_b.3.bn.num_batches_tracked", "lateral.0.cfem_b.4.bn.num_batches_tracked", "lateral.0.ConvLinear.bn.num_batches_tracked", "lateral.0.shortcut.bn.num_batches_tracked", "lateral.1.cfem_a.0.bn.num_batches_tracked", "lateral.1.cfem_a.1.bn.num_batches_tracked", "lateral.1.cfem_a.2.bn.num_batches_tracked", "lateral.1.cfem_a.3.bn.num_batches_tracked", "lateral.1.cfem_a.4.bn.num_batches_tracked", "lateral.1.cfem_b.0.bn.num_batches_tracked", "lateral.1.cfem_b.1.bn.num_batches_tracked", "lateral.1.cfem_b.2.bn.num_batches_tracked", "lateral.1.cfem_b.3.bn.num_batches_tracked", "lateral.1.cfem_b.4.bn.num_batches_tracked", "lateral.1.ConvLinear.bn.num_batches_tracked", "lateral.1.shortcut.bn.num_batches_tracked".

感觉提供的模型（cfenet300_vgg16_coco.pth）有问题

[liuqinglong110]

您好，我现在看不到代码，方便给个链接吗？

[qijiezhao]

你好，我删掉了repo，因为一个版本正在撰写期刊，但是从repo history还是可以找到之前的代码：https://github.com/qijiezhao/CFENet/tree/c8228dd1a9db82fbb9408232f7c737eb7b134537 Best, Qijie Zhao Visual Data Interpreting and Generation Lab, Institute of Computer Science Technology (VDIG.ICST) Peking University 原始邮件 发件人:[email protected] 收件人:qijiezhao/[email protected] 抄送:[email protected]; [email protected] 发送时间:2019年1月10日(周四) 10:24 主题:Re: [qijiezhao/CFENet] 测试报错 (#23) 您好，我现在看不到代码，方便给个链接吗？ — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

[liuqinglong110]

好的，非常感谢！发自我的华为手机-------- 原始邮件 --------主题：Re: [qijiezhao/CFENet] 测试报错 (#23)发件人：QijieZhao 收件人：qijiezhao/CFENet 抄送：liuqinglong110 ,Comment 你好，我删掉了repo，因为一个版本正在撰写期刊，但是从repo history还是可以找到之前的代码：https://github.com/qijiezhao/CFENet/tree/c8228dd1a9db82fbb9408232f7c737eb7b134537 Best, Qijie Zhao Visual Data Interpreting and Generation Lab, Institute of Computer Science Technology (VDIG.ICST) Peking University 原始邮件 发件人:[email protected] 收件人:qijiezhao/[email protected] 抄送:[email protected]; [email protected] 发送时间:2019年1月10日(周四) 10:24 主题:Re: [qijiezhao/CFENet] 测试报错 (#23) 您好，我现在看不到代码，方便给个链接吗？ — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread. —You are receiving this because you commented.Reply to this email directly, view it on GitHub, or mute the thread.