glow particle

这是一个粒子光效的lora模型，主要是做粒子效果，人物方面的效果会比较接近，但其他的，比如动物，水果，建筑的效果不是那么好。触发词glow particle，关键词建议写上 glow particle,translucent fluorescent,optical flares,

使用建议

● 尺寸建议 512x512、512x768、768x768、768x1024、1024x1024

● 大模型推荐使用真实感大模型，写实大模型，角色/人物大模型。

● VAE建议使用 vae-ft-mse-840000-ema-pruned，Clip skip:2，CFG scale：5~8，lora weight：0.5~1

● 采样器：DPM++ SDE Karras、DPM++ 2M Karras、Euler a；迭代步数：20~40 

● 放大算法：Latent、ESRGAN_4x，放大倍数：1.5~2，重绘幅度：0.5~0.8

出人物的时候，一定要开ADtailer，否则脸部扭曲，除人物以外的效果有时候可以，有时候不行，就需要多尝试。

因为是这个风格的第一个版本，还会有一些问题，后续会继续优化。

希望多出一些图，提一些小建议，拜托了！！！

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

This is a lora model of particle light effect, mainly to do particle effect, the effect of the character will be closer, but the effect of other, such as animals, fruits, buildings is not so good. trigger words is glow particle, suggested keywords glow particle,translucent fluorescent,optical flares,

Use suggestion

● Recommended size 512x512, 512x768, 768x768, 768x1024, 1024x1024

● It is recommended to use realistic large models, realistic large models, and role/character large models.

● VAE is recommended to use vae- FT-MSE-84000-EMA-Pruned,

Clip skip:2, CFG scale: 5~8, lora weight: 0.5~1

● Sampler: DPM++ SDE Karras, DPM++ 2M Karras, Euler a;

Number of iteration steps: 20~40

● Amplification algorithm: Latent, ESRGAN_4x, amplification: 1.5~2, redraw range: 0.5~0.8

When the character is out, be sure to open ADtailer, otherwise the face is distorted, the effect other than the character is sometimes OK, sometimes not, you need to try more.

Because it is the first version of this style, there will be some problems, and it will continue to be optimized in the future.

Hope more pictures, some suggestions, please!!