详情
相關作品
v1.0
SeekWest

SeekWest

4.7K
75
223
#風格
#幻想
# 西部
#style lora
#CowGirls
#cowboys
...查看更多

Hi again!

Welcome to "SeekWest". SeekWest it's part of my brand new "Seek" series where I tend to a more complete LORA.

The MAIN purpose of this LORA is to create character arts in an specific style (cowboy, cowgirl and related) but without the needs of specify the clothing (the style will drag to it) and at the same time, being able to specify if the user wants, so it will work both sides. It will act as Lazy LORA (due to how was trained) and it can act as detail LORA the moment you start long-prompting (where you can say what items in the wearing, etc.).

It can create interiors, landscapes and exteriors (I will explain later this since some models are better for this purpose).

- Token is: w3st.

How to use this?

- LORA's strenght recommended: 0.7-1.1 (you can work with 1 with no problems but for advanced prompting or complex things maybe 0.9 is a more safe value).

Note: In the case you don't know you can decide to NOT use the token when playing with values and still it's going to work (however I'd recommend to use it in normal circumstances).

- For characters if you want to use the Lazy Lora tech. "w3st male/female portrait/side portrait" and that should be all. If you want to be more specific with places, outfits, etc. I suggest you to drop the LORA to values between 0.8-0.9 and if you find some problems with specific clothing, 0.7.


- For landscapes/interiors/exteriors I suggest you use Colorful model (I used 3.0), and Dreamshaper (I used 3.32 ClipFix). Others can work but didnt test every single one.

- In terms of general aesthetics if you want some cool anime generations Ghostmix and DarkSushiMix are your friends. If you want overall Colorful v3.0 works very good (same as Dreamshaper or Faetastic).


- If you're thinking on guns and weapons, use ControlNet. AI itself can't get that in an easy way (probably other LORA's for that task can help).

This is LORA is good for:

- Character designing, portraiture. in a range of semi realistic-anime-illustration and probably realistic with RV2.0

- Mixing with other character LORA's.

- Structures, landscapes, interiors, exteriors.

- Popular people stylish-adaptation

This LORA is NOT designed for:

- Extremely complex landscaping/architecture

- Shotgun, gun, weapons holding gens. without ControlNet

Last but not least I will be throwing a Patreon version with almost the same but with some enhancements (although I really think this version is already very good). Indeed a pain in the to train this but happy with the outcome.

IMPORTANT!

I wrote an article here and it's like my finetunning diary for the ones interested with releases upcoming, trainings done, exclusive patreon models previews, personal thoughts, etc. So if you want to see what I'm planning or what I'm into, feel free to take a look.


Feel free to support us at:
https://www.patreon.com/user?u=90661954

Thank you all!.

David.

查看译文

評分與評論

--/5

尚未收到足夠的評分或評論

更多
no-data
暫無數據
與模型對話
公告
2024-04-02
发布模型
2023-06-14
更新模型資訊
模型详情
類型
LORA
发布時間
2023-06-14
基础模型
SD 1.5
触发词
復制
版本介绍

Version 1.1

许可范围
來源: civitai

1. 轉载模型僅供學習與交流分享,其版權及最終解释權归原作者。

2. 模型原作者如需認領模型,請通過官方渠道联系海藝AI工作人員進行認證。我們致力於保護每一位創作者的權益。 點擊去認領

創作许可范围
在線生圖
進行融合
允许下载
商業许可范围
生成圖片可出售或用於商業目的
允许模型轉售或融合后出售
QR Code
下載SeaArt App
在移動端继續你的AI創作之旅