RuntimeError expected input... to have 28 channels, but got 27 channels instead

[ialhashim]

I am getting this error on my own test data (with task 008_VRT_videodenoising_DAVIS)

RuntimeError: Given groups=1, weight of size [96, 28, 1, 3, 3], expected input[1, 27, 32, 128, 128] to have 28 channels, but got 27 channels instead

Full stack:
File "C:\Dev\VRT\models\network_vrt.py", line 1395, in forward x = self.conv_first(x.transpose(1, 2)) File "C:\tools\miniconda3\envs\pt\lib\site-packages\torch\nn\modules\module.py", line 1102, in _call_impl return forward_call(*input, **kwargs) File "C:\tools\miniconda3\envs\pt\lib\site-packages\torch\nn\modules\conv.py", line 590, in forward return self._conv_forward(input, self.weight, self.bias) File "C:\tools\miniconda3\envs\pt\lib\site-packages\torch\nn\modules\conv.py", line 585, in _conv_forward return F.conv3d( RuntimeError: Given groups=1, weight of size [96, 28, 1, 3, 3], expected input[1, 27, 32, 128, 128] to have 28 channels, but got 27 channels instead

[JingyunLiang]

We add noises in the network forward() function. Make sure nonblind_denoising == True here:

VRT/models/network_vrt.py

Line 1376 in c033a79

if self.nonblind_denoising:

[ialhashim]

I see. So this network is trained on a specific kind of noise correct?

[JingyunLiang]

Yes. Following FastDVDNet and PaCNet, we train non-blind denoising models for fair quantitative comparison.

[Gyudori]

So, do we need 4 channel input image for testing denoising?
And if then, where can I get noise-level-map(4th channel) from?

[JingyunLiang]

The input is 3 channel. We add noise-level-map as following:

We add noises in the network forward() function. Make sure nonblind_denoising == True here:

VRT/models/network_vrt.py

Line 1376 in c033a79

if self.nonblind_denoising:

[WEIZHIHONG720]

I meet new problem:deform_conv2d() takes from 3 to 7 positional arguments but 8 were given !\ @JingyunLiang @Gyudori