this is a total experiment :°D i've trained a bunch of pictures of a Fiat 500, it "" "kinda" "" works, generate a lot of crap and doesn't suit very well with illustration style, but something good comes out 😅 Edit : v2 version works a little well with illustration style, integrates better with characters lora, always deformed and out-of-proportions, but give a nice aesthetics to the picture (and personally I like the European city style it gives to the image 😁) but I have still to learn how to train vehicles lora 😂

Files with SHA256: 192a092c61c0bbd85c9f4a00007c9f4f71fba89d8a8c03707e352b2d8f0b40ee