What resolution does inpainting "only masked" use?

[Jonseed]

It used to be that when inpainting you could set the resolution to use for the inpainted area (only masked). But now the UI just shows "resize to" and "resize by" resolutions, which seem more appropriate for img2img tasks. What resolution is it using for inpainting? How is that set?

[w-e-w]

What resolution is it using for inpainting? How is that set?

it used to be that when inpainting you could set the resolution to use for the inpainted area (only masked)

it hasn't changed. it still use whatever every resolution you have set
resize to is equivalent to width and height in txt2img
resize by just a diffrent way of setting width and height, it multiplies your current image size by a factor and use that as the width and height

whatever you have set is the size that it does the stable diffusion generation at

what only masked dose is that it calculate the smallest box that can envelope your mask area
then increase the region on all sides by the size of the padding
then increase the region again to the ratio of your set width and height

then it performs img2img at width and height (the value you set in the ui)

after that it puts the generated composite on to the original image