Test

[01SHENyankun]

I have finished training, but I have encountered problems during the test. I would like to ask what to input at the terminal for test.

[01SHENyankun]

I tested it according to the command of readme and got the following results

Then, it is modified according to the above error report, and the following results appear

[jingyuanli001]

can you try to make the mask a single channel image?

[01SHENyankun]

can you try to make the mask a single channel image?

Hello, can you give me a contact information, my qq2635505974

[01SHENyankun]

I have some questions for you

[su161021]

Hello, may I ask how this problem was finally solved? I also met the same problem, thank you!

[lgl1227]

I have the same question. I want to slove this problem.
Traceback (most recent call last):
File "run.py", line 38, in
run()
File "run.py", line 30, in run
model.test(dataloader, args.result_save_path)
File "E:\PycharmProject\RFR-Inpainting\model.py", line 90, in test
fake_B, mask = self.G(masked_images, masks)
File "E:\Anaconda3\lib\site-packages\torch\nn\modules\module.py", line 1110, in _call_impl
return forward_call(*input, **kwargs)
File "E:\PycharmProject\RFR-Inpainting\modules\RFRNet.py", line 146, in forward
x1, m1 = self.Pconv1(in_image, mask)
File "E:\Anaconda3\lib\site-packages\torch\nn\modules\module.py", line 1110, in _call_impl
return forward_call(*input, **kwargs)
File "E:\PycharmProject\RFR-Inpainting\modules\partialconv2d.py", line 55, in forward
self.update_mask = F.conv2d(mask, self.weight_maskUpdater, bias=None, stride=self.stride, padding=self.padding, dilation=self.dilation, groups=1)
RuntimeError: Given groups=1, weight of size [64, 3, 7, 7], expected input[1, 9, 256, 256] to have 3 channels, but got 9 channels instead

[lgl1227]

I understand. Because the bit depth of these mask images is not consistent, some are 8 bits, some are 24 bits

[lgl1227]

×
√

[01SHENyankun]

× √

你好，请问您这个代码，你进行测试了嘛，有测试指标嘛

[lgl1227]

这个代码我只是跑通了，训练出了模型，但没有进行进一步的测试，没有测试指标。 From: 01SHENyankun Date: 2022-04-15 11:03 To: jingyuanli001/RFR-Inpainting CC: Alex Valentine; Comment Subject: Re: [jingyuanli001/RFR-Inpainting] Test (Issue #60) × √ 你好，请问您这个代码，你进行测试了嘛，有测试指标嘛 ― Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: ***@***.***>

[Jamie-61]

你好，请问你这个问题解决了嘛

我遇到了同样的问题，可不可以请教一下

[lgl1227]

不好意思，你的图片看不到

[Jamie-61]

/home/qsh/anaconda3/bin/python3.7 /home/qsh/ccc/RFR-Inpainting-master/run.py
Model moved to cuda

测试代码中mask只能是外部加载的mask数据集吗（mask_mode只能为2吗），如果mask_mode使用0或1，就会出现上述错误

[lgl1227]

我的理解是mask_mode的0或者1模式 不是用于测试的吧。

[Jamie-61]

嗯嗯！谢谢啦！

[R1ckLou]

I have the same question. I want to slove this problem. Traceback (most recent call last): File "run.py", line 38, in run() File "run.py", line 30, in run model.test(dataloader, args.result_save_path) File "E:\PycharmProject\RFR-Inpainting\model.py", line 90, in test fake_B, mask = self.G(masked_images, masks) File "E:\Anaconda3\lib\site-packages\torch\nn\modules\module.py", line 1110, in _call_impl return forward_call(*input, **kwargs) File "E:\PycharmProject\RFR-Inpainting\modules\RFRNet.py", line 146, in forward x1, m1 = self.Pconv1(in_image, mask) File "E:\Anaconda3\lib\site-packages\torch\nn\modules\module.py", line 1110, in _call_impl return forward_call(*input, **kwargs) File "E:\PycharmProject\RFR-Inpainting\modules\partialconv2d.py", line 55, in forward self.update_mask = F.conv2d(mask, self.weight_maskUpdater, bias=None, stride=self.stride, padding=self.padding, dilation=self.dilation, groups=1) RuntimeError: Given groups=1, weight of size [64, 3, 7, 7], expected input[1, 9, 256, 256] to have 3 channels, but got 9 channels instead

我也是遇到这个问题，请问怎么解决？ 我的mask图片位深度都是相同的8

[youarenotaloneor]

RuntimeError: Given groups=1, weight of size [64, 3, 7, 7], expected input[1, 9, 256, 256] to have 3 channels, but got 9 channels instead
@jingyuanli001 ：I have the above errors in the test. How can I solve them？

[xiaoyanghuha]

It is necessary to use a mask image with a bit depth of 1. I successfully solved this problem and tested the results，If you also want to solve this problem, please contact me “xiaoyanghuha@163.com”

[xiximelon]

comment this line!
# masks = torch.cat([masks]*3, dim = 1)