>>106372515
idk use ComfyUI? Maybe it's the prompt. Keep in mind CFG is literally a scaled diff between inference of the positive and negative prompt. Imagine the negative as a bad picture that has to be coherently spelled out and not pocket spaghetti of bad words like SDXL. This workflow is native and https://files.catbox.moe/znpgqx.png
>boomerized prompt: This is a photo of a fit 19 year old, beautiful, cute, and sexy South American Aztec woman standing in a dark forest at night. She has long, straight black hair with blunt bangs and brown skin, along with large breasts and wide hips. She is wearing a minimal, earthy-brown strapless bandeau top that covers her breasts and a matching loincloth that hangs from her hips, revealing her midriff, abdomen, and thighs. Holding a large, bloody battle axe, she is looking directly at the camera. The forest background consists of tall, dark pine trees with grass covering the ground. The image evokes a dark, moody atmosphere, emphasized by a deep blue and black color palette that highlights the nighttime setting.