Tried my hand at cars. I used the Juggernaut model. The prompt uses regional prompter extension and looks like this:

at night city, summer, sweaty ADDCOMM
RAW photo, Nikon Z6 II Mirrorless Camera, hyper realism, extremely detailed, 8k uhd, dslr, soft lighting, high quality, film grain ADDBASE
ADDROW
ADDCOL
Aston Martin zeekars, hotrod, LED, lora:zeekars:.7 ADDROW
Negative prompt: JuggernautNegative-neg, transition of shapes, blurry, (((numbers on door))), duplicates, close range angle, crop, illustration, drawing, painting, sketching, render, artwork, 3d, cgi, logo, text, letterbox, 3D, render, video game, anime, cartoon, sketch, caption, subtitle, signature, watermark, username, artist name

I use SD upscale and ControlNet Tile for upscaling with 4x_NMKD-Siax_200k. zeekars is a Lora.

I hope SDXL will improve in all the small details for mechanical, geometrical forms. At first, the image looks great but when you look closely, panels separation, headlight, tyres look unrealistic...We will see tomorrow!

The reference family of preprocessors in ControlNet allow you to quickly transfer a style from an image in txt2img or img2img mode. I will demonstrate how to use it in txt2img.

• download this image. I chose the 853x1280 size. We will use it as our ControlNet Reference image.

• setup your ControlNet tab like this.
  ControlNet setup
  If you wonder what is the difference between the reference preprocessors, you can find more info here

-add a prompt close to the style of your image. Very important, set the CFG to a low value (3,4). Otherwise, your result will look very dark and overcooked. This is my settings:
text2img settings

Try to keep the aspect ratio of the original image but do not hesitate to experiment.

• Results:
  original image
  Albert
  Scarlett

AI Art News & Culture by the people, for the people. Click to read Stable Digest, by Stable Foundation, a Substack publication with tens of thousands of readers.