Feature Request Negative Masking

[FlowDownTheRiver]

Feature Request Negative Masking

[FlowDownTheRiver]

I couldn't write enough in the previous post even it was buggy with 0.6 and fixed in 0.7 but I wanted to thank you.I have tested it and the behaviour is like expected.This is a very good utility saving a lot of time than masking manually.

As I python beginner I am trying to learn stuff hoping to reach at some level that I can write some code which people can benefit like yours one day.That is why I envy your work a lot and appreciate it.

Btw having an option to be changed by the user for the delimeter is great and yeah I somehow missed "inpaint not masked" that it can be used as inverting.

This tool definitely has be part of the automatic's repo by default.

[ThereforeGames]

Hi @FlowDownTheRiver,

I like the idea of a negative mask prompt a lot - it has been added in 0.0.6! Thanks for the suggestion.

I have changed the default delimeter from "|" to "," cos it makes more sense and easy to write.I would like to see it as a default.

The vertical pipe was selected as the delimiter because I don't want to interfere with common characters like "," in case the user wishes to use such characters in a single prompt search. I do think "|" will yield better results most of the time, but maybe not always.

That said, you're welcome to change this to anything you prefer (it is assigned to a variable called delimiter_string in the latest version) and I may look into having some sort of user preferences file if I end up developing this script much further.

In addition to these a checkbox for inverting the mask can be a cool feature I believe.

If you're on the "Inpaint a part of image" tab of the UI, there is an option to "Inpaint not mask" which effectively inverts the txt2mask script.

[FlowDownTheRiver]

Getting value error

'mask': <PIL.PngImagePlugin.PngImageFile image mode=RGBA size=2048x3072 at 0x2645BF6DDB0>}, None, 0, 20, 0, 4, 0, False, False, 1, 1, 1, 13, 0.6, -1.0, -1.0, 0, 0, 0, False, 512, 512, 0, 0, 64, True, 0, 8, '', '', '', '', 1, 50, 0, 4, 1, '

Recommended settings: Sampling Steps: 80-100, Sampler: Euler a, Denoising strength: 0.8

', 128, 4, ['left', 'right', 'up', 'down'], 1, 0.05, 128, 4, 0, ['left', 'right', 'up', 'down'], False, None, '', False, 'cloth,neck', 'face', 96, True, 0, '

If you like my work, please consider showing your support on Patreon. Thank you! ❤

', 1, '', 4, '', True) {}
Traceback (most recent call last):
File "C:\stable-diffusion-webui\modules\ui.py", line 135, in f
res = list(func(*args, **kwargs))
File "C:\stable-diffusion-webui\webui.py", line 58, in f
res = func(*args, **kwargs)
File "C:\stable-diffusion-webui\modules\img2img.py", line 134, in img2img
processed = modules.scripts.scripts_img2img.run(p, *args)
File "C:\stable-diffusion-webui\modules\scripts.py", line 159, in run
processed = script.run(p, *script_args)
File "C:\stable-diffusion-webui\scripts\txt2mask.py", line 156, in run
processed = processing.process_images(p)
File "C:\stable-diffusion-webui\modules\processing.py", line 312, in process_images
p.init(all_prompts, all_seeds, all_subseeds)
File "C:\stable-diffusion-webui\modules\processing.py", line 534, in init
mask = mask.crop(crop_region)
File "C:\stable-diffusion-webui\venv\lib\site-packages\PIL\Image.py", line 1171, in crop
raise ValueError("Coordinate 'right' is less than 'left'")
ValueError: Coordinate 'right' is less than 'left'

[ThereforeGames]

Hmm, seems to be related to using input images that are not 512x512 - looking into it.

[ThereforeGames]

Okay, should be fixed now. I switched to another image inversion function. No clue why the previous one was spazzing out though.