海凪コウ/Koh (Minagi Kou) Style - Anima Base

This is my first LoRA training. After 4 iterations of adjusting the dataset and parameters, I got this decent version. It is designed for the Anima base model. I trained with 100 images from the artist's manga. If you have any questions, please feel free to leave a message.

Two versions are provided: one trained with AdamW optimizer and one with Prodigy optimizer. In my limited testing, both versions perform similarly. You can download and try them yourself.

Usage:

- Trigger word: koh_style

- Recommended LoRA weight: 0.7 – 1.0

- Recommended CFG: 3.5 – 5

- Recommended steps: 24 – 36

- Recommended sampler: euler_ancestral / er_sde

Disclaimer:

This model is for personal, non-commercial, and learning purposes only. Do not sell generated images.

---

海凪コウ/Koh (Minagi Kou) 画风 - Anima 底模

这是我第一次训练 LoRA，迭代了 4 个版本，通过调整数据集和参数，得到这个还算可以的版本。模型为 Anima 底模设计。我使用了 100 张图进行训练，训练集来自该作者的漫画。有什么问题可以给我留言。

提供了两个版本：一个使用 AdamW 优化器，一个使用 Prodigy 优化器。在我有限的测试中，两个版本效果差不多，可以自行下载尝试。

使用建议：

- 触发词： koh_style

- 推荐权重： 0.7 – 1.0

- 推荐 CFG： 3.5 – 5

- 推荐步数： 24 – 36

- 推荐采样器： euler_ancestral / er_sde

免责声明：

本模型仅供个人学习、测试和非商业用途，请勿售卖生成图片。