Model Description by Creator

Checkpoint Type: SDXL 1.0/Corruption/Chaos/Test for merged models and LoRAs

Follow me on Twitter: @YamerOfficial Discord: yamer_ai

If you have enjoyed this checkpoint:


[Before creation comes Destruction.]

From the abyss of Hell, a checkpoint emerges; An embodiment of annihilation and turmoil. Crafted in the essence of destruction, its sole purpose is to obliterate every image, intentionally designed to be the epitome of corruption and devastation. This model shall bear the name "Hakai", the technique used by the gods of destruction.

This is a very special model, it goes against all principles of a good SDXL checkpoint, in fact it tries to be the worst while still generating images that are recognizable to the human eye, the model is based on the "Unstable Diffusers V11 Hakai" (not released yet).


True to its name, the outcomes produced by this checkpoint inherently reflect chaos and corruption. Its design is deliberate, with the sole intent of embodying corruption.

This model serves as an ideal testing ground for future merged models and experimenting with LoRAs. This checkpoint is great for testing purposes because of the unique style it has that is not matched by any other model on Civitai.


HOW TO USE (MUST READ):

This model is hard to use without the recommended settings, that is the reason why you must read these instructions before generating any images!

  1. Use of low Hires steps may result in lower quality (All the showcase images used 50 hires steps)

  2. Use of denoise strength can change everything in the final image, low denoise strength for less unique images, higher denoise strength between 0.3 and 0.8 to generate a better unique look

  3. A low CFG will generate images with almost no detail, use a CFG above 4

  4. Steps are very important, I recommend the use of at least 10 steps

Here are some examples:

Best result, the main figure is recognizable; Denoising strength: 0.7, Hires upscale: 2.5, Hires steps: 50, Hires upscaler: 4x_foolhardy_Remacri

Denoising strength: 0.3, Hires upscale: 2.5, Hires steps: 1, Hires upscaler: 4x_foolhardy_Remacri

Denoising strength: 0.15, Hires upscale: 2.5, Hires upscaler: 4x_foolhardy_Remacri

Denoising strength: 0.15, Hires upscale: 2.5, Hires steps: 1, Hires upscaler: 4x_foolhardy_Remacri

These are the settings I used to generate the images in the showcase, if you want to reproduce the images I generated follow these settings:

The Hires steps are set to zero because A1111 will automatically match the steps you have (50) and do the same amount for Hires steps (50).

To achieve this style a SD 1.5 VAE was used.


Unstable Diffuser and Stable Destruction comparison:

(The third model is Unstable V11 ヤメールの帝国, unfortunately XYZ plot isn't capable of recognizing japanese characters)

You can find the images full resolution here: https://civitai.com/user/Yamer/posts


Creator note:

Hi, I'm an Italian creator on a mission to spread the joy of using AI to generate images and the progress it has made in the past few years. I craft checkpoints and prompts that anyone can use to generate their own art and share it. My ultimate aim is to be a good creator, sharing the models I create as a hobby with others.


*-* It works best with:

  • LoRA: No LoRA will save you from its destruction

  • Restore Face: You can try, you will fail to save the face

  • Clip skip: 1-2 (The model works well with clip skip set to both.)

  • SDXL Refiner: Not used with my models!

1) Reality is often mischievous, create your own.
2) Imagination will take you a million light years from home. But will it bring you back?
3) Be chaotic, be creative, never stop improving and never put a limit to yourself.

    -Yamer Grimlock

𝕮𝖍𝖊𝖈𝖐𝖕𝖔𝖎𝖓𝖙 𝕽𝖆𝖙𝖎𝖓𝖌:

  • Quality: 1/10

  • Stability (Variety): 9/10, the lower = the more variety with one prompt

  • Flexibility: 3/10

  • CFG resistance: 30/30 (no matter the CFG it will produce chaos)

  • Style strength: 10/10

  • Style Variety: 1/10

Images Generated With This Model

Loading...

Train Models

Train AI Models On Your Own Pictures

You can use your own pictures to train AI models. Reimagine yourself in new ways with our AI-powered image generator.
Create pictures of yourself around the world, in your dream outfit, as a video game character, or in countless art styles. The possibilities are endless.
Train Your Own Model

Generate AI Images. Simple and Easy

AIEasyPic is the #1 AI image generator. Generate images with simple prompts, swap faces, train models to reimagine yourself, and more.