|
--- |
|
tags: |
|
- text-to-image |
|
- stable-diffusion |
|
- diffusion-models-class |
|
- dreambooth-hackathon |
|
- wildcard |
|
--- |
|
|
|
# If you want to use this style , the prompt: wgz style |
|
**I trianed this weight by using dreambooth** |
|
There are three models in the library, V1, V2, and beta |
|
V1:WOLP.ckpt |
|
V2:wgz v2.ckpt |
|
BEAT :wgz beta.ckpt |
|
|
|
|
|
Let me introduce these versions , |
|
|
|
V1 is the earliest. use 512 resolution for training, |
|
|
|
Advantages: the generation logic is good, text2img is used well |
|
|
|
Disadvantages: There will be less screen details and less imitation of strokes. The whites of the eyes turn red easily. |
|
|
|
V2 adjusted the training set and used 768 resolution training |
|
|
|
Advantages: more details, better facial portrayal |
|
|
|
Disadvantages: Eyeballs tend to generate green, and the logic of text generation is not good, resulting in poor usability of text2img. |
|
|
|
BETA: A version of the training map that is relatively complex has poor logic, but sometimes satisfactory results can be obtained in the img2 img. Try if you want! |
|
|
|
**v2** |
|
Next, I will show the nice output of V2, which is made by using the img2img. I will also give the configuration I use. |
|
|
|
**parameters;** |
|
|
|
Prompt :wgz style,portrait of a beautiful women highly detailed,perfect femine face,Classical oil painting,by masamune shirow, by William-tae Kim |
|
|
|
Negative prompt: old,men,two poeple,Three dimensional facial features,deep shadows on the five senses,tall nose,Eye shadow,[High bridge], [narrow nose],Deep orbital fossa |
|
|
|
Steps: 20, Sampler: DPM++ 2S a Karras, CFG scale: 9.5, Denoising strength: 0.51, Mask blur: 4 |
|
I hope you guys enjoy this model and create works that satisfied |
|
|
|
|
|
 |
|
**BETA** |
|
|
|
Below I will show the nice output of a beta. I like its abstract color swatches, but the production rate is not high |
|
|
|
 |
|
|
|
I hope you guys enjoy this model and create works that satisfied |
|
|
|
|
|
Below is the introduction about V1,and some nice output by using V1 |
|
**v1** |
|
|
|
-------------------------- |
|
|
|
Training with the Elden ring Style ckpt （SD1.5） |
|
# link：[eldenring-ckpt](https://huggingface.co/nitrosocke/elden-ring-diffusion) |
|
# Training parameters： |
|
steps： 12500 stpes |
|
Learning rate：1e-6 |
|
int images number：about 100 |
|
class image number ：200 |
|
|
|
|
|
**There are some samples:** |
|
 |
|
 |
|
|
|
PS:I try hard to achieve the painting style I want by using stable diffusion. I have tried a lot, with failures and successes. With my current technology, I can hardly find progress now. So this model will not be updated in a short time. If you have ideas or want to improve together, you can contact me,happy new year guys. |
|
|
|
Telegram: https://t.me/hhhyxnh |
|
QQ:1602821649 |
|
|
|
** More pictures in the image folder |
|
|
|
thank you. |
