|
--- |
|
|
|
|
|
{} |
|
--- |
|
|
|
# True Buruma |
|
|
|
<!-- Provide a quick summary of what the model is/does. --> |
|
Buruma(not bloomers) is a world heritage. |
|
 |
|
|
|
## Updates |
|
|
|
- **buruma-230327_20**: Latest test for renewal 2.0 version. |
|
- **buruma20renew_test-230323**: Test for renewal 2.0 version.Activeted taraining tags. |
|
|
|
|
|
## Features |
|
A LoRA file for StableDiffusion-webui. |
|
|
|
It was traind with photographs of ancient burumas, not imitaions, |
|
especially close-up shots. |
|
It will give us highly-detailed buruma images. |
|
|
|
This model is an experimental. It will be improved for your support. see Donations |
|
## How to Use |
|
|
|
Recommended LoRA weight is 0.3. |
|
```prompt example |
|
<lora:buruma_gym_uniform_TrueBuruma10:0.3>, buruma, gym_uniform |
|
``` |
|
```prompt example |
|
Negative Prompt: monochrome |
|
``` |
|
**Notes:** This model is **Double Trigger**.Both of "buruma" and "gym_uniform" are nessesary. |
|
|
|
If you want red burumas, |
|
```prompt example |
|
red_buruma, gym_uniform |
|
``` |
|
But it's hard to draw.Please wait a verious clolored version. |
|
|
|
Clothes Lifting looks good... |
|
```prompt example |
|
buruma, gym_uniform, clothes_lift, nipples |
|
``` |
|
### LoRA Block Weight Plugin Tequinic |
|
|
|
Disable **OUT04**, you can reduce the upper_body effect. |
|
So these are can be(but a little bit hard.I can't support it) |
|
- Wearing another upper wear |
|
- Getting more huge breasts |
|
- Coexsisting with another LoRAs |
|
|
|
```prompt example |
|
<lora:buruma_gym_uniform_TrueBuruma10:0.3:1,1,1,1,1,1,1,1,1,0,1,1,1,1,1,1,1> |
|
``` |
|
See example and prompts. https://majinai.art/ja/i/MwIql8F |
|
|
|
### Recommended SD models |
|
- Anything V3(I had confirmed good outputs) |
|
- BloodOrangeMix(I love it) https://huggingface.co/WarriorMama777/OrangeMixs |
|
- **Realistic models also available** |
|
|
|
### LoRA Training Parameters |
|
<!-- Provide a longer summary of what this model is. --> |
|
- **Max Resolution:** 512,512 |
|
- **Dimeisions:** 16 |
|
- **Network Alpha:** 8 |
|
- **Learning Rate:** 0.0001 |
|
- **Epoch:** 20 |
|
- **Training Based Model:** ACertainty https://huggingface.co/JosephusCheung/ACertainty |
|
- **Source Pictures:** 83 max |
|
- **Calculating Graphic Card:** RTX2060(6GB) It's so hard! |
|
|
|
## Term of Use |
|
|
|
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. --> |
|
- **Not for corporate activities**(Give me money) |
|
- **Prohibit redistribution** |
|
|
|
|
|
### Donate me if you feel like that your life will be good |
|
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. --> |
|
Thank you for 6 cup of coffee!(Mar.21,2023) |
|
|
|
- Coffee to me: https://ko-fi.com/kanaejun |
|
- BitCoin: 3H33ahQnpL84S4kSKWxCPzKMoEjtdgMBRy |
|
- LiteCoin: MWYA5K8jNo9AoCadvPpoMHs9bCEB23QYuk |
|
|
|
## Author |
|
|
|
kanaejun, Japan |
|
- https://twitter.com/kanaejun |
|
|
|
|
|
|
