README.md

---
license: creativeml-openrail-m
pipeline_tag: text-to-image
tags:
  - stable-diffusion
  - stable-diffusion-diffusers
  - text-to-image
library_name: diffusers
---
<style>
* {
  box-sizing: border-box;
}

.column {
  float: left;
  width: 100%;
  padding: 5px;
}

.column2 {
  float: left;
  width: 50%;
  padding: 5px;
}

.column3 {
  float: left;
  width: 33.33%;
  padding: 5px;
}

.row::after {
  content: "";
  clear: both;
  display: table;
}

@media screen and (max-width: 400px) {
  .column {
    width: 100%;
  }
  .column2 {
    width: 100%;
  }
}
</style>

<center><img src=https://huggingface.co/mekabu/MagicalMix_v2/resolve/main/MagicalMix_v2_logo.png></center>

<center><p><h3>What's this model</h></p></center>

Based on <a href="https://huggingface.co/mekabu/MagicalMix_v1">MagicalMix v1</a>, we aimed for a softer picture.

By samplers, both anime and softer (like illustration) styles can be expressed.

<center><p><h3>Setting and examples</h></p></center>

**Embeddings**

<a href="https://huggingface.co/datasets/gsdf/EasyNegative">EasyNegative</a>

<a href="https://huggingface.co/datasets/Nerfgun3/bad_prompt">bad_prompt_version2</a>

VAE:<a href="https://huggingface.co/andite/pastel-mix/blob/main/pastel-waifu-diffusion.vae.pt">pastel-waifu-diffusion.vae.pt</a>

**[Softer]**
```
Sampler: Euler a
Steps: 40
Hires fix: x2
Upscaler: Latent
Hires steps: 30
Denoising strength: 0.5
CFG scale: 8
Clip skip: 2
```

<div class=row>
<div class=column2><center><img src=https://huggingface.co/mekabu/MagicalMix_v2/resolve/main/Sample/SampleA_01.png></center></div>
<div class=column2><center><img src=https://huggingface.co/mekabu/MagicalMix_v2/resolve/main/Sample/SampleA_02.png></center></div>
<div class=column2><center><img src=https://huggingface.co/mekabu/MagicalMix_v2/resolve/main/Sample/SampleA_03.png></center></div>
</div>

**[Anime]**
```
Sampler: DPM++ 2M Karras
Steps: 25
Hires fix: x2
Upscaler: R-ESRGAN 4x++ Anime6B
Hires steps: 20
Denoising strength: 0.5
CFG scale: 8
Clip skip: 2
```

<div class=row>
<div class=column2><center><img src=https://huggingface.co/mekabu/MagicalMix_v2/resolve/main/Sample/SampleB_01.png></center></div>
<div class=column2><center><img src=https://huggingface.co/mekabu/MagicalMix_v2/resolve/main/Sample/SampleB_02.png></center></div>
<div class=column2><center><img src=https://huggingface.co/mekabu/MagicalMix_v2/resolve/main/Sample/SampleB_03.png></center></div>
</div>

<br>

\*\***Any other good presets out there, let me know!****

<center><p><h3>Model recipe</h></p></center>

|No.|Interpolation Method|Model_A|Model_B|Tertiary Model|Output|
|:---:|:---:|:---:|:---:|:---:|:---:|
|1|Weighted sum @ 0.45|Basil_mix_fixed|uenomigiDB_v1|-|Bronze1|
|2|Weighted sum @ 0.7|Bronze1|ligneClaireAnime_v1|-|Bronze2|
|3|Weighted sum @ 0.6|Bronze2|abysshell_ver3|-|Bronze3|
|4|Add Difference @ 0.75|Bronze3|dalcefoHitoria_v3|EasyNegative|**Silver**|

|Model_A|Model_B|Weight|Output|
|:---:|:---:|:---:|:---:|
|MagicalMix_v1|Silver|SMOOTHSTEP|MagicalMix_v2|