Azazeal Spectra

This model has the TCD (Trajectory Consistency Distillation) Lora merged into it.

The model is capable to generate images at 2 steps with CFG set at 1.1 to 1.3 but I do recommend using the settings as set below in Usage in ComfyUI. Unfortunately I dont use any other UI other from ComfyUI, but the settings should be the same for other UI's.

Positive prompt:

masterpiece, highly detailed, "your prompt, subject, etc", intricate details, detailed background

Negative prompt:

nsfw, topless, nude, worst quality, bad quality, text, watermark, logo, blurry, extra fingers, missing fingers, naked, nude

Usage in ComfyUI:

CFG: 1.3

Steps: 3

Sampler: DPMPP-SDE

Scheduler: Karras, sgm-uniform and AYS-SDXL

DO NOT USE ACCELERATING LORAS (Such as Lightning,, etc) WITH THIS MODEL AS YOUR IMAGES WILL BE OVER SATURATED AND BURNED.

Don't use DPM++ 2M/ DPM++ 2M Karras as Sampler! It won't work good!

Don't use this to harm others... Please use this responsible! You are fully responsible for your actions!

And most important:

Don't use this to create CP!