|
--- |
|
license: other |
|
license_name: faipl-1.0-sd |
|
license_link: https://freedevproject.org/faipl-1.0-sd/ |
|
pipeline_tag: text-to-image |
|
base_model: |
|
- Laxhar/noobai-XL-Vpred-1.0 |
|
tags: |
|
- stable-diffusion |
|
- stable-diffusion-xl |
|
--- |
|
|
|
# Abydos_Noob_v-pred-1.0.0 |
|
|
|
|
|
 |
|
|
|
Modified NoobAI-XL(v-prediction) with Blue Archive style |
|
|
|
[Civitai model page](https://civitai.com/models/923120) |
|
|
|
## Prompt Guidelines |
|
Almost same as the base model |
|
|
|
## Recommended Prompt |
|
None(Works good without `masterpiece, best quality`) |
|
|
|
## Recommended Negative Prompt |
|
`worst quality, bad quality, lowres, photoshop \(medium\), abstract` |
|
|
|
To improve the quality of background, add `simple background, transparent background` to Negative Prompt. |
|
|
|
## Recommended Settings |
|
Steps: 12-24 |
|
|
|
Sampler: DPM++ 2M(dpmpp_2m) |
|
|
|
Scheduler: Simple or SGM Uniform |
|
|
|
Guidance Scale: 2-5 |
|
|
|
### Hires.fix |
|
Upscaler: 4x-UltraSharp or Latent(nearest-exact) |
|
|
|
Denoising strength: 0.5(0.6-0.7 for latent) |
|
|
|
## Training steps |
|
1. Make 2 models from NoobAI(=A,B), A with ZTSNR, B w/o ZTSNR |
|
2. Merge A and B MBW(0,0,0,0,0,0.5,0.5,0,0.7,0.7,0.7,0.7,0.7,0.7,0.5,0.5,0,0,0,0) Adjust(0,0,0,0,-0.05,0,0,0)=tmp1 |
|
3. tmp1 + spo_sdxl_10ep_4k-data_lora_webui x 1 + sdxl-boldline x -0.25 = Result |
|
|
|
## Training scripts: |
|
[sd-scripts](https://github.com/kohya-ss/sd-scripts) |
|
|
|
## Notice |
|
This model is licensed under [Fair AI Public License 1.0-SD](https://freedevproject.org/faipl-1.0-sd/) |
|
|
|
If you make modify this model, you must share both your changes and the original license. |
|
|
|
You are prohibited from monetizing any close-sourced fine-tuned / merged model, which disallows the public from accessing the model's source code / weights and its usages. |
