Sign In

【Illustration 】Juaner-Ghibli-Style Hayao Miyazaki style animated movie Flux

253
1.3k
124
Verified:
SafeTensor
Type
Checkpoint Merge
Stats
597
Reviews
Published
Dec 29, 2024
Base Model
Flux.1 D
Hash
AutoV2
12BAF979B7
The FLUX.1 [dev] Model is licensed by Black Forest Labs. Inc. under the FLUX.1 [dev] Non-Commercial License. Copyright Black Forest Labs. Inc.
IN NO EVENT SHALL BLACK FOREST LABS, INC. BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH USE OF THIS MODEL.

Update Tip:

This v2.0 version has been enhanced in the style stability and pic quality.

Description of an AI game:

My friend Tianyou and I created an AI challenge game: https://www.fluxultimate.com/turing-test . come and try it.

This game mainly about whether people can tell a painting is generated by AI or an artist within a specified time. Very interesting. We also created a global ranking.

Hi, I'm Juaner.

This is a large model that combines the Flux dev model with 7 different art style LoRAs that I've trained myself.

This model is particularly good at generating illustrations in the style of Hayao Miyazaki, but it has not learned character names, so if you want to generate specific characters, you need to combine it with character LoRAs.

In this version, I tested a total of ten recipes, and this one is currently the recipe I'm most satisfied with. I'm sharing it with you all to play around with. I hope you all have fun with it.

Recommended prompt ideas:

You can directly describe the subject and the contents of the scene using natural language descriptions. You can also add some anime illustration-related prompts to stabilize the style (during testing, I found that some special prompts, such as emphasizing a photographic style in the prompt, could cause the model to lose the illustration style).

Welcome to share any interesting images you generate! ღ( ´・ᴗ・` )

嗨,我是卷儿

这是一个使用flux dev模型和我自己训练的7个画风lora融合成的大模型。

这个模型比较擅长生成宫崎骏风格的插画,但是没有学习角色名字,所以如果想要出固定角色的话需要搭配角色lora来使用。

这一版本一共测试了十种配方,这个配方是我目前比较满意的一版,分享出来给大家玩儿,祝大家玩儿得开心。

推荐的提示词思路:

直接描述主体和场景里面的内容即可,使用自然语言细致描述。也可以加上一些动漫插画相关的提示词来使风格更加稳定(测试中发现:部分特殊的提示词,比如prompt中过于强调摄影风格的时候,会失去插画风格)

欢迎大家分享你生成的有趣的图片,ღ( ´・ᴗ・` )比心