Particalize I2V Lora

If you like my lora, you can come and discuss via discord channel

可以通过 QQ 3877209711 联系我

It is a Lora that can Particalize anything. It might require a little bit of imagination to prompt this Lora. Posts are welcome.

这是一个可以将物体粒子化的Lora。写提示词的时候可能需要一些想象力。欢迎使用完之后返视频到post。

---------------------------------------------------------------------------------------------------------------------------------

Please use version1.5, Suggested strength: 0.4-1.0！！！

Suggested step: 30 cfg: 6.0. For better result, DO NOT use Teacache, it will do great damage to the Lora model. Turning on the slg of Kijai workflow might increase the quality.

请使用最新版本v1.5，建议强度0.4-1.0！！！想要效果好，不要开Teacache,否则生成质量会很差！然后最好开kj的slg

建议步数30 CFG: 6.0

---------------------------------------------------------------------------------------------------------------------------------

Prompt examples: something is disintegrating or turning into something ("something" need to be substituted)

eg(weight 1.0). The yellow Pikachu is disintegrating or turning into small yellow leaves. Then a Nintendo Switch appears on the sofa.

eg(weight 0.45). A figurine wearing a yellow jacket and black shirt underneath. It is disintegrating or turning into many white bread slices.

If you wanna do small particles use weight 1.0, if you wanna be more creative use weight 0.4-0.5 (see my post that a figurine turning to bread slices)

You can use "background change" as the negative prompt.

提示词：什么 is disintegrating or turning into 什么 （“什么”要根据你的需求替换）

例如（权重1.0）：The yellow Pikachu is disintegrating or turning into small yellow leaves. Then a Nintendo Switch appears on the sofa.

例如（权重0.4-0.5）：A figurine wearing a yellow jacket and black shirt underneath. It is disintegrating or turning into many white bread slices.

如果你想做颗粒化那就用权重1.0，如果你想更加有创意那么就用权重0.4-0.5（看我的post那个变成小面包的dio）

你可以把“background change”作为反向提示词

---------------------------------------------------------------------------------------------------------------------------------

Trained on WAN2.1-I2V-480p model, 720p-i2v model might work as well.

在WAN2.1-I2V-480p模型上训练的，但是WAN2.1-I2V-720p模型应该也可以用