Trained it in like 15 minutes with only 50 images to see if it could work and it kinda does work but i've only tested with blood orange mix and it will probably will look bad most of the times, i'll probably try to remake this with more pics and with better tagging to see if something decent can be done in the meantime feel free to play with it. You need to use "a photo of a car, from the side, aniwrap" for starters then you can describe the girl, you probably also have to additionally add "car, vehicle focus, photo background" somewhere else on the prompt. You can merge it with character lora's but will have to play with the weights, you can also do from the front instead of from the side but it doesnt really work that well as most of the pics were from the side.

Files with SHA256: b1c1f977a10134019484abe75d4e1e55f25a9485fec56157b557e3dbaa74dbe8