File size: 3,136 Bytes
6d44eac e7e5177 6d44eac e7e5177 6d44eac 5a7c9ef 6d44eac e7e5177 5a7c9ef e7e5177 6d44eac 198856e e7e5177 198856e 6d44eac 198856e 6d44eac 9b0c536 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 |
---
datasets:
- KBlueLeaf/danbooru2023-sqlite
language:
- en
library_name: diffusers
---
# Kohaku XL Gamma
A SDXL anime base model aims to create unique artworks.
<div style="display:flex; flex-direction: row; width: 100%;">
<div><img style="display:block;" src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/afc1dc39-3a1c-4db9-9942-0b2c150124ce/original=true/00946.jpeg"/></div>
<div><img style="display:block;" src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/20751b43-3d91-4802-b5cb-ce3857d1027a/original=true/00937.jpeg"/></div>
<div><img style="display:block;" src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/fade32c7-7cb8-4a69-84f3-d3da0f4b319d/original=true/00943.jpeg"/></div>
</div>
---
## Introduction
This model can be seen as a derivative of Animagine XL 3.0 project.<br>
Basically I'm collaborating with Linaqruf for making better Anime base model (and it is obvious that we have different goal/target)<br>
We share our models and technique to improve our models' quality.<br>
And that is also how this model been created.<br>
### Base7
Kohaku-XL base7 is resumed from beta7 and use same dataset that beta series have used. But this time I use my own metadata system to create captions. (Can be taken as advanced version of what linaqruf used, will open source it soon)
Trainin details:<br>
LR: 8e-6/2e-6<br>
Scheduler: constant with warmup<br>
Batch size: 128 (batch size 4 * grad acc 16 * gpu count 2) <br>
### Gamma rev1
Kohaku-XL Gamma rev1 is a merged model which combine the learned diff from anxl3 and kohaku xl base 7. With this forumla:
gamma rev1 = beta7 + 0.8 * (anxl3 - anxl2) + 0.5 * (base7 - beta7)
---
## Usage
#### Parameter
This model is trained under 768x1024 to 1024x1024 ARB. It is recommended to use pixel count within 786432 ~ 1310720.<br>
Recommended CFG scale is 4~7.<br>
Sampler should not matter.<br>
#### Tagging
This model use my own system for quality tags or something like that. <br>
So although this model combine the diff weight from anxl3, I will still recommend user to use mine (or both) tagging system.<br>
The format of prompt is as same as anxl3. (You can check the sample images I post)
Rating tags:
* General: safe
* Sensitive: sensitive
* Questionable: nsfw
* Explicit: explicit, nsfw
Quality tags (Better to worse):
* Masterpiece
* best quality
* great quality
* good quality
* normal quality
* low quality
* worst quality
Year tags (New to Old):
* newest
* recent
* mid
* early
* old
You may meet some subtle mosaic-like artifact, that may be caused by high-lr or bad resizing/image encoding.<br>
I will try to fix it in next version. For now, try to use R-ESRGAN anime6b or SCUNet models for fixing it.<br>
---
## Future plan
Since my dataset have some resize/webp artifacts that will harm the models. I will recreate my dataset based on my new system (and opensource it once I done it).
The next plan is to train model on larger (3M-6M) dataset with better configuration (which will require A100s and I plan to spend about 2000-10000 USD on it, if you like my works, consider to sponsor me via buy-me-a-coffee or some BTC-sutff)
|