anima_pencil-XL / anima_pencil-XL-v1.5.0-recipe.md
bluepen5805's picture
Create anima_pencil-XL-v1.5.0-recipe.md
da38e99 verified
|
raw
history blame
2.16 kB
- [AUTOMATIC1111/stable-diffusion-webui](https://github.com/AUTOMATIC1111/stable-diffusion-webui)
- [hako-mikan/sd-webui-supermerger](https://github.com/hako-mikan/sd-webui-supermerger)
# Checkpoints
- anima_pencil-XL-v1.0.0
- [blue_pencil-XL-v4.0.1](https://huggingface.co/bluepen5805/blue_pencil-XL)
- [ANIMAGINE XL 3.0](https://huggingface.co/cagliostrolab/animagine-xl-3.0)
- [ANIMAGINE XL 2.0](https://huggingface.co/Linaqruf/animagine-xl-2.0)
- [AAM XL Anime Mix v1.0](https://civitai.com/models/269232?modelVersionId=303526)
- [/Anime/bulldozer_BETA v2.0](https://civitai.com/models/264323?modelVersionId=315336)
- [AingDiffusion XL v0.4](https://civitai.com/models/226403?modelVersionId=304293)
- [Anime Illust Diffusion XL v0.71](https://civitai.com/models/124189?modelVersionId=318932)
- [SD-XL 1.0-base](https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0)
# LoRAs
- self-made QualitySlider ([Concept Sliders](https://github.com/rohitgandikota/sliders))
- [lcm-lora-sdxl](https://huggingface.co/latent-consistency/lcm-lora-sdxl)
# Recipe
1. blue_pencil-XL-v4.0.1 + ( ANIMAGINE XL 3.0 - ANIMAGINE XL 2.0 ) x 1.0, Add difference(trainDifference)
2. ( [1] x 1.0 ) + ( anima_pencil-XL-v1.0.0 x 0 ), Weight sum(cosineB)
3. ( [2] x 0.85 ) + ( blue_pencil-XL-v4.0.1 x 0.15 ), Weight sum(normal), Elemental: "BASE:conditioner.embedders.0.transformer.text_model.*":0.6
4. [3] + ( AAM XL Anime Mix v1.0 - SD-XL 1.0-base ) x 0.46, Add difference(smoothAdd)
5. [4] + ( /Anime/bulldozer_BETA v2.0 - ANIMAGINE XL 3.0) x 1.0, Add difference(smoothAdd)
6. [5] + ( AingDiffusion XL v0.4 - ANIMAGINE XL 3.0) x 0.3, Add difference(smoothAdd)
7. [6] + ( Anime Illust Diffusion XL v0.71 - ANIMAGINE XL 3.0) x 0.35, Add difference(smoothAdd)
8. ( [7] x 1.0 ) + ( [2] x 0.15 ), Weight sum(normal), Elemental: "BASE:conditioner.embedders.0.transformer.text_model.*":0.8
9. [8] + QualitySlider x 0.65
10. [9] + adjust(0,2,1,1,0.5,0,0,0)
# Recipe(lcm)
1. SD-XL 1.0-base + lcm-lora-sdxl:1.0
2. anima_pencil-XL-v1.5.0 + ( [1] - SD-XL 1.0-base ) x 0.5, Add difference(trainDifference)
3. ( [2] x 0.7 ) + ( anima_pencil-XL-v1.0.0 x 0.3 ), Weight sum(cosineA)