AI Images

I use ComfyUI to generate images according to exacting specifications. For the redheaded model below in a blue dress I used two prompts, one with a LoRA and one without. I did the same for the blonde model surrounded by balls. A sample prompt is below:

Positive prompt:

cinematic film still portrait of a red headed fashion model with blue eyeshadow and blue eyeliner wearing a couture gown in the style of a high fashion photo shoot, bright colors

Negative prompt:

text, watermark

LoRA used: Juggernaut Cinematic XL

Control after generate: fixed

Sampler name: dpmpp_2s_ancestral

Steps: 35

CFG: 7.0

Another sample prompt:

Positive prompt:

asian high fashion model dressed in houndstooth suit with red heels posing against a graffiti wall in Wynwood Miami

negative prompt:

text, watermark

Juggernaut Cinematic XL LoRA

Strength 1.0

Control after generate: fixed

Steps 35

CFG 8.0

Sampler name: dpmpp_2s_ancestral

Denoise: 1.00

With ComfyUI, the key is to use a good model, which can be downloaded from CivitAI.com. I prefer cinematic/photorealistic models. I sometimes use LoRAs to then tweak a model's output or to change something fundamental about a particular generation.