GUI and operation I would like to see these improvements and suggestions #127
Description
I suggest the following improvements:
- Zoom Function - using the webbrowser zoom doesnt work very well on 16:9 images
- Show the used seed number when using -1 so that its possible to reuse the seed of the last generation
- Stable Diffusion saves the prompt, seed number and settings into the PNG file header and also gives automatically filenames so that I dont have to give all my files myself names. These infos in the PNG file header would be very useful so that you always know with which settings this particular image was made.
- The painted mask should be possible to be saved !! Its really annoying having to redraw the whole inpaint as soon as you reopen the program :-/
- Allow uploading inpaint masks. Addons for Stable Diffusion like "Segment Anything" have auto detection for subjects and can create automatically already a subject mask. Let us be able to use it. Most of the time you dont have to do anything more because its mostly pretty accurate. And if - the most of the work is already done by the mask then. or maybe add even yourself a subject detection?
- Why it stores all my Stable Diffusion Models another time in the Magicquill folder with pinokio? So I have now all my models twice on the system due to this. And there are even the SDXL models copied over, which it doesnt even support, coming to the next point
- Add support for more modern models - SDXL etc. They most likely will do better results.
- Maybe you manage to add a reference image feature. So that we can say: Take subject from picture 1 and replace it with the one at picture 2.