Error The size of tensor a (4) must match the size of tensor b (3)

[demiguelm]

HI! After update to new A1, some error appears :(

Works with the regular non-fixed postprocessing_pixelization.py
Everything with the color/brightness fix was working great but...

initialize network with normal
*** Error completing request
*** Arguments: ('task(iz8svjo899oawb6)', 0.0, <PIL.Image.Image image mode=RGBA size=1280x720 at 0x2152D7A5BA0>, None, '', '', True, True, 0.0, 4, 0.0, 512, 512, True, 'None', 'None', 0, False, 1, False, 1, 0, False, 0.5, 0.2, False, 0.9, 0.15, 0.5, False, False, 384, 768, 4096, 409600, 'Maximize area', 0.1, False, ['Horizontal'], False, ['Deepbooru'], False, False, 4, True, False, 4, True, True, True, 15, 1, None, False, '0', '0', 'inswapper_128.onnx', 'CodeFormer', 1, True, 'None', 1, 1, 1, 0, 0, 0.5, 'CUDA', False, 0, 'None', '', None, False, False, 0.5, 0, 'tab_single') {}
Traceback (most recent call last):
File "F:\SD1.9\webui\modules\call_queue.py", line 57, in f
res = list(func(*args, **kwargs))
File "F:\SD1.9\webui\modules\call_queue.py", line 36, in f
res = func(*args, **kwargs)
File "F:\SD1.9\webui\modules\postprocessing.py", line 131, in run_postprocessing_webui
return run_postprocessing(*args, **kwargs)
File "F:\SD1.9\webui\modules\postprocessing.py", line 71, in run_postprocessing
scripts.scripts_postproc.run(initial_pp, args)
File "F:\SD1.9\webui\modules\scripts_postprocessing.py", line 198, in run
script.process(single_image, **process_args)
File "F:\SD1.9\webui\extensions\stable-diffusion-webui-pixelization\scripts\postprocessing_pixelization.py", line 272, in process
in_t = process(pp.image).to(devices.device)
File "F:\SD1.9\webui\extensions\stable-diffusion-webui-pixelization\scripts\postprocessing_pixelization.py", line 155, in process
return trans(img)[None, :, :, :]
File "F:\SD1.9\system\python\lib\site-packages\torchvision\transforms\transforms.py", line 95, in call
img = t(img)
File "F:\SD1.9\system\python\lib\site-packages\torch\nn\modules\module.py", line 1518, in _wrapped_call_impl
return self.call_impl(*args, **kwargs)
File "F:\SD1.9\system\python\lib\site-packages\torch\nn\modules\module.py", line 1527, in call_impl
return forward_call(*args, **kwargs)
File "F:\SD1.9\system\python\lib\site-packages\torchvision\transforms\transforms.py", line 277, in forward
return F.normalize(tensor, self.mean, self.std, self.inplace)
File "F:\SD1.9\system\python\lib\site-packages\torchvision\transforms\functional.py", line 363, in normalize
return F_t.normalize(tensor, mean=mean, std=std, inplace=inplace)
File "F:\SD1.9\system\python\lib\site-packages\torchvision\transforms_functional_tensor.py", line 928, in normalize
return tensor.sub(mean).div(std)
RuntimeError: The size of tensor a (4) must match the size of tensor b (3) at non-singleton dimension 0

---

[Taruu]

can u show file postprocessing_pixelization.py?
from script folder. or more precisely the code on line 199

[Taruu]

Or is everything already working?

[demiguelm]

Yeah, the problem is the fix file: #18

I updated the code so it will also accept 0.75 or 2.25 values from slider , also you have ratio to restore shadows and highlights, defaults work ok but youre free to change them depending on image postprocessing_pixelization.zip

I manage to fix adding line 266.

[Taruu]

you can also try this changes from @y4my4my4m pull request #31
I check later. I think this solution more better