I don't really know why this happens, tensor a (1460) must match the size of tensor b (1464) at non-singleton dimension 3

chaiNNer-org / chaiNNer

A node-based image processing GUI aimed at making chaining image processing tasks easy and customizable. Born as an AI upscaling application, chaiNNer has grown into an extremely flexible and powerful programmatic image processing application.

https://chaiNNer.app

GNU General Public License v3.0

4.47k stars 280 forks source link

I don't really know why this happens, tensor a (1460) must match the size of tensor b (1464) at non-singleton dimension 3 #2989

Open Tulip-0333 opened 2 months ago

Tulip-0333 commented 2 months ago

Information:

Chainner version: v0.24.1
OS: Windows 11

Description When using two images with this layout*, it brings up the error `tensor a (1460) must match the size of tensor b (1464) at non-singleton dimension 3*, it seems to be any* image with the same dimensions, though I don't really know.

* **

Logs main.log No other logs

joeyballentine commented 2 months ago

Can you try using the latest nightly? I believe we've fixed this and it just hasn't been released yet

Tulip-0333 commented 2 months ago

Not fixed..

joeyballentine commented 2 months ago

Ok, then this is just something we have to fix. In the meantime though, you should be able to get around it by giving one of the images a little bit of padding.

Lukeyy19 commented 1 month ago

I am getting the same issue but I don't understand how padding helps? If I add padding to either the image or the mask then I just get an error saying the image and mask must have the same resolution, but without the padding I get the same "The size of tensor a (1482) must match the size of tensor b (1488) at non-singleton dimension 3" despite both the image and the mask having a width of 1482...