Battlefield Key Visual AIO SDXL

This model is based on SDXL, trained according to the key art of the Battlefield game series. This model integrates the artistic styles from various Battlefield titles and can be invoked on demand. It aims to provide a high-contrast style similar to original concept arts.

How to use:

Users need to have the SDXL base model (or fine-tuned SDXL model) as a prerequisite. Then, invoke this model in prompts. The recommended weight is 0.7~1.0. Use bf3, bf4, bf1, bfv as prompts to specify a particular conceptual style.

When using specific concepts, I suggest using the following prompts together:

• bf3: long exposure, tracer

• bf4: bokeh, depth of field

• bf1: light streak, lens flare

• bfv: light streak, lens flare, bokeh

• Overall: explosion, fire, fire sparks, burning

Limitations:

• According to my personal testing, this model performs poorly on base models which are finetuned in anime styles.

• It may randomly generate some distorted military uniforms/tacticals/firearms.

I am still working on optimizing this model. If you have feedback or better training ideas, please contact me. Thank you.

战地系列主视觉图风格SDXL版

本模型是基于SDXL作为基础模型，根据游戏《战地》系列的概念美术进行训练的，旨在提供类似其概念美术的高对比度风格。其中包括了多款《战地》作品中的不同美术要素。

如何使用：

用户需要拥有SDXL基础模型（或者微调过的SDXL模型）作为前置，然后在Prompt里调用此模型。推荐的权重为0.7~1.0。 用bf3, bf4, bf1, bfv作为Prompts来指定特定作品中的概念风格。

在使用特定的概念时，我建议配合以下Prompt一并使用：

• bf3: long exposure, tracer

• bf4: bokeh, depth of field

• bf1: light streak, lens flare

• bfv: light streak, lens flare, bokeh

• 总体：explosion, fire, fire sparks, burning
  

局限性：

• 根据我本人的测试，该模型在动漫风格特化的底模上表现欠佳。

• 可能会随机出现一些扭曲的军服/防弹衣/枪械。

我还在致力于优化本模型，如果有建议或者更好的训练思路，请联系我，感谢。