These are early days for SketchUp Diffusion so you all have a unique opportunity to shape this tool in ways that make sense for your specific workflows.
This thread is a place to share information about features that would add value to your workflows, improve performance, and otherwise make the feature more valuable or delightful to use.
Please don’t be shy. Let us know what else you want Diffusion to do for you.
I added the icon to my topbar and closed the toolbar. yet whenever I launch SU, the toolbar reopens again and again.
I understand this is something the extension dev has to adjust, I have a couple other extensions doing the same, and the devs never answered, could be nice not to have another
On my Mac System when opening the styles option the window has no scrollbar and does not respond to my mouse scroll wheel or two finger drag. Depending on the size of the window one might not realized there were more than 6 options unless they resized the window.
The released version lacks some features that were there in earlier stages, like creation of seamless textures and the ability to view the generated image in a separate window. I found these quite useful.
The available “hand-rendered” styles produce, IMO, an effect that has perhaps a touch that looks both a bit too personal and at the same time, somewhat amateurish. I don’t know if it would be possible to feed the AI with images from the great architectural watercolorists of the past.
Hopefully the distant-enough past to avoid legal or even moral copyright and intellectual-property infringement concerns. (Not a dig at @Anssi - a dig at for-profit companies that make money in generative AI on the uncompensated fruits of other peoples’ creative labor.)
I meant that. Joseph Gandy comes to mind. He was perhaps the greatest architectural illustrator of all time, working for John Soane and other British Regency architects. He died 180 years ago. His work, too, is perhaps too personal but certainly not lackluster. Not looking for imitations of him
They do. A workaround is to export a 2D image from SketchUp in ortho view. Open a fresh Sketchup file and import that image into Sketchup as an image, position the camera so it is looking straight at the image (use a standard view) then apply the Diffusion generate. Diffusion will still recognize the distinct elements and apply styles to the image.
After I save a scene is there a way to see what prompt I used for it?
If I generate an image, than orbit the camera a bit, but forget to refresh the scene in sketchup diffusion the generated images are off position… and I cannot correct it through the scenes because camera location in the scene pane is grayed out for some reason.