wcy-cs / fishfsrnet Goto Github PK

您好，在运行fsr文件夹中main_parsing.py文件时，遇到下列问题：
TypeError: conv2d() received an invalid combination of arguments - got (Tensor, Parameter, Parameter, tuple, tuple, tuple, int), but expected one of:

• (Tensor input, Tensor weight, Tensor bias, tuple of ints stride, tuple of ints padding, tuple of ints dilation, int groups)
  didn't match because some of the arguments have invalid types: (Tensor, Parameter, Parameter, tuple, tuple, tuple, int)
• (Tensor input, Tensor weight, Tensor bias, tuple of ints stride, str padding, tuple of ints dilation, int groups)
  didn't match because some of the arguments have invalid types: (Tensor, Parameter, Parameter, tuple, tuple, tuple, int)
  请问是net结构的问题吗？应该怎么解决？期盼您的回复，感谢！

您好，请问训练集是自己先预处理了Celeba，将其大小调整为128×128作为ground truth HR人脸，然后将ground truth的大小调整为64×64, 32×32和16×16作为相应的LR人脸，这个是分别放在不同的文件夹下并命名为LR等了吗？但好像与程序中的scale == 4，8,16对不太上。非常期待您的回答，谢谢！

作者您好！请问parsingnet和fishfsrnet在训练阶段都用的是CelebA中的168854张图像吗？另外想问一下两个训练阶段采用的epoch数目和batch size分别是多少呀！十分感谢

请问完整的训练与测试代码会更新吗？

作者您好！我最近用您公布的源代码重新训练了FishFSRNet×4和FishFSRNet×8网络，在做测试的时候，×8的模型没有任何问题可以得出PSNR值，但是×4的模型不知为何有如下的报错，请问该如何解决呢？十分感谢！
Traceback (most recent call last):
File "D:\pythonProject1\FishFSRNet-main\fsr\test.py", line 41, in
main()
File "D:\pythonProject1\FishFSRNet-main\fsr\test.py", line 20, in main
net.load_state_dict(pretrained_dict)
File "D:\Anaconda\envs\fish\Lib\site-packages\torch\nn\modules\module.py", line 2153, in load_state_dict
raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format(
RuntimeError: Error(s) in loading state_dict for FISHNET:
Missing key(s) in state_dict: "refine2.0.refine8.conv.0.body.0.weight", "refine2.0.refine8.conv.0.body.0.bias", "refine2.0.refine8.conv.0.body.2.weight", "refine2.0.refine8.conv.0.body.2.bias", "refine2.0.refine8.conv.1.body.0.weight", "refine2.0.refine8.conv.1.body.0.bias", "refine2.0.refine8.conv.1.body.2.weight", "refine2.0.refine8.conv.1.body.2.bias", "refine2.0.attention.mlp.1.weight", "refine2.0.attention.mlp.1.bias", "refine2.0.attention.mlp.3.weight", "refine2.0.attention.mlp.3.bias", "refine2.1.refine8.conv.0.body.0.weight", "refine2.1.refine8.conv.0.body.0.bias", "refine2.1.refine8.conv.0.body.2.weight", "refine2.1.refine8.conv.0.body.2.bias", "refine2.1.refine8.conv.1.body.0.weight", "refine2.1.refine8.conv.1.body.0.bias", "refine2.1.refine8.conv.1.body.2.weight", "refine2.1.refine8.conv.1.body.2.bias", "refine2.1.attention.mlp.1.weight", "refine2.1.attention.mlp.1.bias", "refine2.1.attention.mlp.3.weight", "refine2.1.attention.mlp.3.bias", "refine2.2.down2.0.weight", "refine2.2.down2.0.bias", "refine2.2.down2.2.weight", "refine2.2.down2.2.bias", "refine2.2.refine8.conv.0.body.0.weight", "refine2.2.refine8.conv.0.body.0.bias", "refine2.2.refine8.conv.0.body.2.weight", "refine2.2.refine8.conv.0.body.2.bias", "refine2.2.refine8.conv.1.body.0.weight", "refine2.2.refine8.conv.1.body.0.bias", "refine2.2.refine8.conv.1.body.2.weight", "refine2.2.refine8.conv.1.body.2.bias", "refine2.2.attention.mlp.1.weight", "refine2.2.attention.mlp.1.bias", "refine2.2.attention.mlp.3.weight", "refine2.2.attention.mlp.3.bias", "refine2.3.down1.weight", "refine2.3.down1.bias", "refine2.3.down2.0.weight", "refine2.3.down2.0.bias", "refine2.3.down2.2.weight", "refine2.3.down2.2.bias", "refine2.3.refine8.conv.0.body.0.weight", "refine2.3.refine8.conv.0.body.0.bias", "refine2.3.refine8.conv.0.body.2.weight", "refine2.3.refine8.conv.0.body.2.bias", "refine2.3.refine8.conv.1.body.0.weight", "refine2.3.refine8.conv.1.body.0.bias", "refine2.3.refine8.conv.1.body.2.weight", "refine2.3.refine8.conv.1.body.2.bias", "refine2.3.attention.mlp.1.weight", "refine2.3.attention.mlp.1.bias", "refine2.3.attention.mlp.3.weight", "refine2.3.attention.mlp.3.bias", "refine2.4.down1.weight", "refine2.4.down1.bias", "refine2.4.refine2.conv.0.body.0.weight", "refine2.4.refine2.conv.0.body.0.bias", "refine2.4.refine2.conv.0.body.2.weight", "refine2.4.refine2.conv.0.body.2.bias", "refine2.4.refine2.conv.1.body.0.weight", "refine2.4.refine2.conv.1.body.0.bias", "refine2.4.refine2.conv.1.body.2.weight", "refine2.4.refine2.conv.1.body.2.bias", "refine2.4.refine4.conv.0.body.0.weight", "refine2.4.refine4.conv.0.body.0.bias", "refine2.4.refine4.conv.0.body.2.weight", "refine2.4.refine4.conv.0.body.2.bias", "refine2.4.refine4.conv.1.body.0.weight", "refine2.4.refine4.conv.1.body.0.bias", "refine2.4.refine4.conv.1.body.2.weight", "refine2.4.refine4.conv.1.body.2.bias", "refine2.4.refine8.conv.0.body.0.weight", "refine2.4.refine8.conv.0.body.0.bias", "refine2.4.refine8.conv.0.body.2.weight", "refine2.4.refine8.conv.0.body.2.bias", "refine2.4.refine8.conv.1.body.0.weight", "refine2.4.refine8.conv.1.body.0.bias", "refine2.4.refine8.conv.1.body.2.weight", "refine2.4.refine8.conv.1.body.2.bias", "refine2.4.attention.mlp.1.weight", "refine2.4.attention.mlp.1.bias", "refine2.4.attention.mlp.3.weight", "refine2.4.attention.mlp.3.bias", "refine2.4.conv.weight", "refine2.4.conv.bias", "refine2.5.refine2.conv.0.body.0.weight", "refine2.5.refine2.conv.0.body.0.bias", "refine2.5.refine2.conv.0.body.2.weight", "refine2.5.refine2.conv.0.body.2.bias", "refine2.5.refine2.conv.1.body.0.weight", "refine2.5.refine2.conv.1.body.0.bias", "refine2.5.refine2.conv.1.body.2.weight", "refine2.5.refine2.conv.1.body.2.bias", "refine2.5.refine4.conv.0.body.0.weight", "refine2.5.refine4.conv.0.body.0.bias", "refine2.5.refine4.conv.0.body.2.weight", "refine2.5.refine4.conv.0.body.2.bias", "refine2.5.refine4.conv.1.body.0.weight", "refine2.5.refine4.conv.1.body.0.bias", "refine2.5.refine4.conv.1.body.2.weight", "refine2.5.refine4.conv.1.body.2.bias", "refine2.5.refine8.conv.0.body.0.weight", "refine2.5.refine8.conv.0.body.0.bias", "refine2.5.refine8.conv.0.body.2.weight", "refine2.5.refine8.conv.0.body.2.bias", "refine2.5.refine8.conv.1.body.0.weight", "refine2.5.refine8.conv.1.body.0.bias", "refine2.5.refine8.conv.1.body.2.weight", "refine2.5.refine8.conv.1.body.2.bias", "refine2.5.attention.mlp.1.weight", "refine2.5.attention.mlp.1.bias", "refine2.5.attention.mlp.3.weight", "refine2.5.attention.mlp.3.bias", "refine2.5.conv.weight", "refine2.5.conv.bias", "up1.0.0.weight", "up1.0.0.bias", "up2.0.0.weight", "up2.0.0.bias", "up3.0.0.weight", "up3.0.0.bias", "up_stage3.0.body.0.weight", "up_stage3.0.body.0.bias", "up_stage3.0.body.2.weight", "up_stage3.0.body.2.bias", "up_stage3.0.attention_layer1.spatial_layer1.weight", "up_stage3.0.attention_layer1.spatial_layer1.bias", "up_stage3.0.attention_layer1.spatial_layer3.weight", "up_stage3.0.attention_layer1.spatial_layer3.bias", "up_stage3.0.attention_layer2.mlp.1.weight", "up_stage3.0.attention_layer2.mlp.1.bias", "up_stage3.0.attention_layer2.mlp.3.weight", "up_stage3.0.attention_layer2.mlp.3.bias", "up_stage3.0.conv.weight", "up_stage3.0.conv.bias", "up_stage3.0.conv_feature.0.weight", "up_stage3.0.conv_feature.0.bias", "up_stage3.0.conv_parsing.0.weight", "up_stage3.0.conv_parsing.0.bias", "up_stage3.0.conv_fusion.weight", "up_stage3.0.conv_fusion.bias", "up_stage3.0.attention_fusion.weight", "up_stage3.0.attention_fusion.bias", "up_stage3.1.body.0.weight", "up_stage3.1.body.0.bias", "up_stage3.1.body.2.weight", "up_stage3.1.body.2.bias", "up_stage3.1.attention_layer1.spatial_layer1.weight", "up_stage3.1.attention_layer1.spatial_layer1.bias", "up_stage3.1.attention_layer1.spatial_layer3.weight", "up_stage3.1.attention_layer1.spatial_layer3.bias", "up_stage3.1.attention_layer2.mlp.1.weight", "up_stage3.1.attention_layer2.mlp.1.bias", "up_stage3.1.attention_layer2.mlp.3.weight", "up_stage3.1.attention_layer2.mlp.3.bias", "up_stage3.1.conv.weight", "up_stage3.1.conv.bias", "up_stage3.1.conv_feature.0.weight", "up_stage3.1.conv_feature.0.bias", "up_stage3.1.conv_parsing.0.weight", "up_stage3.1.conv_parsing.0.bias", "up_stage3.1.conv_fusion.weight", "up_stage3.1.conv_fusion.bias", "up_stage3.1.attention_fusion.weight", "up_stage3.1.attention_fusion.bias", "down1.conv.weight", "down1.conv.bias", "down_stage1.0.body.0.weight", "down_stage1.0.body.0.bias", "down_stage1.0.body.2.weight", "down_stage1.0.body.2.bias", "down_stage1.0.attention_layer1.spatial_layer1.weight", "down_stage1.0.attention_layer1.spatial_layer1.bias", "down_stage1.0.attention_layer1.spatial_layer3.weight", "down_stage1.0.attention_layer1.spatial_layer3.bias", "down_stage1.0.attention_layer2.mlp.1.weight", "down_stage1.0.attention_layer2.mlp.1.bias", "down_stage1.0.attention_layer2.mlp.3.weight", "down_stage1.0.attention_layer2.mlp.3.bias", "down_stage1.0.conv.weight", "down_stage1.0.conv.bias", "down_stage1.0.conv_feature.0.weight", "down_stage1.0.conv_feature.0.bias", "down_stage1.0.conv_parsing.0.weight", "down_stage1.0.conv_parsing.0.bias", "down_stage1.0.conv_fusion.weight", "down_stage1.0.conv_fusion.bias", "down_stage1.0.attention_fusion.weight", "down_stage1.0.attention_fusion.bias", "down_stage1.1.body.0.weight", "down_stage1.1.body.0.bias", "down_stage1.1.body.2.weight", "down_stage1.1.body.2.bias", "down_stage1.1.attention_layer1.spatial_layer1.weight", "down_stage1.1.attention_layer1.spatial_layer1.bias", "down_stage1.1.attention_layer1.spatial_layer3.weight", "down_stage1.1.attention_layer1.spatial_layer3.bias", "down_stage1.1.attention_layer2.mlp.1.weight", "down_stage1.1.attention_layer2.mlp.1.bias", "down_stage1.1.attention_layer2.mlp.3.weight", "down_stage1.1.attention_layer2.mlp.3.bias", "down_stage1.1.conv.weight", "down_stage1.1.conv.bias", "down_stage1.1.conv_feature.0.weight", "down_stage1.1.conv_feature.0.bias", "down_stage1.1.conv_parsing.0.weight", "down_stage1.1.conv_parsing.0.bias", "down_stage1.1.conv_fusion.weight", "down_stage1.1.conv_fusion.bias", "down_stage1.1.attention_fusion.weight", "down_stage1.1.attention_fusion.bias", "conv_tail1.weight", "conv_tail1.bias", "conv.weight", "conv.bias", "up21.0.0.weight", "up21.0.0.bias", "conv_tail2.weight", "conv_tail2.bias", "up22.0.0.weight", "up22.0.0.bias", "up23.0.0.weight", "up23.0.0.bias", "conv_tail3.weight", "conv_tail3.bias", "up2_stage3.0.body.0.weight", "up2_stage3.0.body.0.bias", "up2_stage3.0.body.2.weight", "up2_stage3.0.body.2.bias", "up2_stage3.0.attention_layer1.spatial_layer1.weight", "up2_stage3.0.attention_layer1.spatial_layer1.bias", "up2_stage3.0.attention_layer1.spatial_layer3.weight", "up2_stage3.0.attention_layer1.spatial_layer3.bias", "up2_stage3.0.attention_layer2.mlp.1.weight", "up2_stage3.0.attention_layer2.mlp.1.bias", "up2_stage3.0.attention_layer2.mlp.3.weight", "up2_stage3.0.attention_layer2.mlp.3.bias", "up2_stage3.0.conv.weight", "up2_stage3.0.conv.bias", "up2_stage3.0.conv_feature.0.weight", "up2_stage3.0.conv_feature.0.bias", "up2_stage3.0.conv_parsing.0.weight", "up2_stage3.0.conv_parsing.0.bias", "up2_stage3.0.conv_fusion.weight", "up2_stage3.0.conv_fusion.bias", "up2_stage3.0.attention_fusion.weight", "up2_stage3.0.attention_fusion.bias", "up2_stage3.1.body.0.weight", "up2_stage3.1.body.0.bias", "up2_stage3.1.body.2.weight", "up2_stage3.1.body.2.bias", "up2_stage3.1.attention_layer1.spatial_layer1.weight", "up2_stage3.1.attention_layer1.spatial_layer1.bias", "up2_stage3.1.attention_layer1.spatial_layer3.weight", "up2_stage3.1.attention_layer1.spatial_layer3.bias", "up2_stage3.1.attention_layer2.mlp.1.weight", "up2_stage3.1.attention_layer2.mlp.1.bias", "up2_stage3.1.attention_layer2.mlp.3.weight", "up2_stage3.1.attention_layer2.mlp.3.bias", "up2_stage3.1.conv.weight", "up2_stage3.1.conv.bias", "up2_stage3.1.conv_feature.0.weight", "up2_stage3.1.conv_feature.0.bias", "up2_stage3.1.conv_parsing.0.weight", "up2_stage3.1.conv_parsing.0.bias", "up2_stage3.1.conv_fusion.weight", "up2_stage3.1.conv_fusion.bias", "up2_stage3.1.attention_fusion.weight", "up2_stage3.1.attention_fusion.bias".
Unexpected key(s) in state_dict: "refine2.0.attention.body.0.weight", "refine2.0.attention.body.0.bias", "refine2.0.attention.body.2.conv1.weight", "refine2.0.attention.body.2.conv1.bias", "refine2.0.attention.body.2.conv3.weight", "refine2.0.attention.body.2.conv3.bias", "refine2.0.attention.body.2.conv5.weight", "refine2.0.attention.body.2.conv5.bias", "refine2.0.attention.body.2.conv7.weight", "refine2.0.attention.body.2.conv7.bias", "refine2.0.attention.attention_layer2.mlp.1.weight", "refine2.0.attention.attention_layer2.mlp.1.bias", "refine2.0.attention.attention_layer2.mlp.3.weight", "refine2.0.attention.attention_layer2.mlp.3.bias", "refine2.1.attention.body.0.weight", "refine2.1.attention.body.0.bias", "refine2.1.attention.body.2.conv1.weight", "refine2.1.attention.body.2.conv1.bias", "refine2.1.attention.body.2.conv3.weight", "refine2.1.attention.body.2.conv3.bias", "refine2.1.attention.body.2.conv5.weight", "refine2.1.attention.body.2.conv5.bias", "refine2.1.attention.body.2.conv7.weight", "refine2.1.attention.body.2.conv7.bias", "refine2.1.attention.attention_layer2.mlp.1.weight", "refine2.1.attention.attention_layer2.mlp.1.bias", "refine2.1.attention.attention_layer2.mlp.3.weight", "refine2.1.attention.attention_layer2.mlp.3.bias", "refine2.2.attention.body.0.weight", "refine2.2.attention.body.0.bias", "refine2.2.attention.body.2.conv1.weight", "refine2.2.attention.body.2.conv1.bias", "refine2.2.attention.body.2.conv3.weight", "refine2.2.attention.body.2.conv3.bias", "refine2.2.attention.body.2.conv5.weight", "refine2.2.attention.body.2.conv5.bias", "refine2.2.attention.body.2.conv7.weight", "refine2.2.attention.body.2.conv7.bias", "refine2.2.attention.attention_layer2.mlp.1.weight", "refine2.2.attention.attention_layer2.mlp.1.bias", "refine2.2.attention.attention_layer2.mlp.3.weight", "refine2.2.attention.attention_layer2.mlp.3.bias", "refine2.3.attention.body.0.weight", "refine2.3.attention.body.0.bias", "refine2.3.attention.body.2.conv1.weight", "refine2.3.attention.body.2.conv1.bias", "refine2.3.attention.body.2.conv3.weight", "refine2.3.attention.body.2.conv3.bias", "refine2.3.attention.body.2.conv5.weight", "refine2.3.attention.body.2.conv5.bias", "refine2.3.attention.body.2.conv7.weight", "refine2.3.attention.body.2.conv7.bias", "refine2.3.attention.attention_layer2.mlp.1.weight", "refine2.3.attention.attention_layer2.mlp.1.bias", "refine2.3.attention.attention_layer2.mlp.3.weight", "refine2.3.attention.attention_layer2.mlp.3.bias", "up1.body.0.weight", "up1.body.0.bias", "up2.body.0.weight", "up2.body.0.bias", "up21.body.0.weight", "up21.body.0.bias", "up22.body.0.weight", "up22.body.0.bias".
size mismatch for refine2.0.conv.weight: copying a param with shape torch.Size([64, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 192, 1, 1]).
size mismatch for refine2.1.conv.weight: copying a param with shape torch.Size([64, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 192, 1, 1]).
size mismatch for refine2.2.conv.weight: copying a param with shape torch.Size([64, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 192, 1, 1]).
size mismatch for refine2.3.conv.weight: copying a param with shape torch.Size([64, 128, 1, 1]) from checkpoint, the shape in current model is torch.Size([64, 192, 1, 1]).

你好，请问预训练的模型可以提供一下吗

作者您好！可以麻烦您分享一下Helen测试集的GT图像的语义分割图吗？我看之前的链接里没有，十分感谢！！

您好！请问是怎么通过BiSeNet来提取人脸解析图的？相关代码可以提供一下吗？感谢！

File "/FishFSRNet-main/fsr/fishfsrnet.py", line 98, in init
self.reduc = common.channelReduction()
AttributeError: module 'common' has no attribute 'channelReduction'
您好，在运行python main_parsing.py训练整个网络时会出错，显示fishfsrnet.py中没有channelReduction，去到common.py中确实没有找到channelReduction的定义，请问这个common.channelReduction()表示的是什么呢？能否在common.py中添加对于的代码，感谢

请问训练ParsingNet的时候，按照readme给的命令，出现 TypeError: init() missing 1 required positional argument: 'args'
这个报错，参数传入问题，应该如何解决呢？感谢！