Your current browser has compatibility issues. Please switch toChrome for uninterrupted access.

Ghibli Style GPT-4o

1.4k
9
19
7
可商用Anime StyleBoyStyle Boost|Japanese animeGhibli StyleGPT-4o
最近更新:25/04/01首次發布:25/03/28
Anime Style,Boy,Style Boost,LoRA,FLUX.1Image info
Anime Style,Boy,Style Boost,LoRA,FLUX.1Image info
Anime Style,Boy,Style Boost,LoRA,FLUX.1Image info
Anime Style,Boy,Style Boost,LoRA,FLUX.1Image info
Anime Style,Boy,Style Boost,LoRA,FLUX.1Image info
Anime Style,Boy,Style Boost,LoRA,FLUX.1Image info

Ghibli style GPT-4o version

This GPT-4o style is very popular recently, let's take advantage of it

LoRA weight recommendation: around 0.8

Trigger word: Studio ghibli style

Prompt word template: image medium + subject description + background description + other supplements

For example: This is a digital drawing in an anime style featuring two characters standing on a ship's deck during sunset. The background shows a gradient sky transitioning from deep purple to orange. The girl, with long brown hair, wears a blue dress with a white underlayer and a yellow scarf, arms outstretched. The boy, with tousled brown hair, is behind her, embracing her from behind. Both have serene expressions, eyes closed, suggesting a moment of shared tranquility. The ship's rigging is visible, adding to the nautical theme.

You can provide themes and formats for Deepseek to polish and write


The rest can be kept as default, as follows

Negative prompt words: Not required

Base template: F.1 basic algorithm, star stream

Sampling method: Euler

CFG: 3.5

Iteration steps: 30

Resolution: 1024x1024, preferably multiples of 64, such as 1024x1536, 1152x768, 1344x896

Ensure the resolution is around 1024. When the character occupies a small proportion of the screen, structural errors may occur. You can turn on high-resolution repair, redraw the amplitude 0.3-0.6, and enlarge it by 2 times. The higher the amplitude, the more obvious the picture changes


Model Usage Statement:

1. Without authorization, you may not host this model and its derivative versions on other service platforms or integrate them into applications and websites.

2. You may not use this model to create and disseminate illegal or harmful content. Please comply with relevant laws, regulations and public order and good customs.

3. You shall use the content generated based on this model after independent judgment. Any intellectual property issues and legal risks that may arise from this shall be borne by you. I am not responsible for any losses caused by this.

全部展開

討論3

最熱門
|
最新
發送
avatar
Jianfeng Chen
您好,请问该lora什么时候可以开放下载?
2025/04/01 03:30:39
0回覆
avatar
HongkeArtist
回覆@Jianfeng Chensooner or later
2025/04/01 10:56:29
0回覆
avatar
HongkeArtist
Use the comfyUI workflow I2I : https://www.shakker.ai/modelinfo/db3df9bb7b49485a9f984c960f11d9e1
2025/04/01 10:54:45
0回覆
沒有更多評論
Hongke
藝術家
424
1.5k
208.2k
3.2k
收藏
下載
(292.23MB)
已驗證:2025/03/28
v1.0
專屬詳細資訊
類型
LoRA
線上生成次數
1.4k
下載次數
9
基礎算法
FLUX.1
功能描述
GPT-4o version of Ghibli
觸發詞
Studio ghibli style
推薦參數
推薦搭配Checkpoint
Flux.1 Model Family_Black Forest Labs_FLUX.1-dev-fp8null_0005
推薦的重量
0.8
CFG
3.5
VAE
高清放大演算法
R-ESRGAN_4x+ Anime6B
版權
可以線上生成
不可以合併
商業版權商業使用規則
生成圖片可用於商業目的(僅限會員)立即升級 >>
模型不能轉售或作為合併後的模型出售
*版權範圍由創建者自行設定,用戶必須按照要求和標準使用

圖庫33

最熱門
|
最新
avatar
Propriocepti
Image info
avatar
Hongke
Image info
avatar
Hongke
Image info
avatar
Hongke
Image info
avatar
Hongke
Image info
avatar
Hongke
Image info
avatar
Hongke
Image info
avatar
Hongke
Image info
avatar
艺源RAO
Image info
avatar
Hongke
Image info
avatar
Hongke
Image info
avatar
Hongke
Image info
avatar
Hongke
Image info
avatar
Hongke
Image info
avatar
Hongke
Image info
avatar
Hongke
Image info