Moving around with img2img, textual inversion and controlnet
I first made the image square, with a bit of wiggle-room for the model
I first made the image square, with a bit of wiggle-room for the model
this is the image I started with
👉
I explored this a bit
then I added some 3D into the mix by using a textual inversion embedding I had trained on some of my more 3D generations.
from here I seem to have used ControlNet to get back to a flatter version.
I played around in that flat space for a bit.
I played around in that flat space for a bit.
