It would be nice if the mask options could be set in the prompt, along their subprompts. Something like:
((ultra-detailed)), ((illustration)), 2girls <mask:0.2:20>
AND ((ultra-detailed)), ((illustration)), 2girls, black hair <mask:0.8:rect(0,0,0.5,1)>
AND ((ultra-detailed)), ((illustration)), 2girls, blonde hair <mask:0.8:rect(0.5,0,0.5,1)>
The syntax would be <mask:weight[:endAt]:rect(x,y,w,h)>. The endAt option would only be needed in one of the masks (usually the full image one) and it should allow a percentage over the steps value, and in the rect part using numbers <=1 would use percentages of the image size. Not setting a rect would imply a full image. This format would also allow future expansion for other kinds of masks (circles, etc.), assuming that it is possible.
This would also make unnecessary all the settings (except the enable checkbox and the visualize button, which is very useful).
It would be nice if the mask options could be set in the prompt, along their subprompts. Something like:
The syntax would be
<mask:weight[:endAt]:rect(x,y,w,h)>
. The endAt option would only be needed in one of the masks (usually the full image one) and it should allow a percentage over the steps value, and in the rect part using numbers <=1 would use percentages of the image size. Not setting a rect would imply a full image. This format would also allow future expansion for other kinds of masks (circles, etc.), assuming that it is possible.This would also make unnecessary all the settings (except the enable checkbox and the visualize button, which is very useful).