Trained on: NAI, works on both anime and real/3D checkpoints (note: checkpoints that are 100% photorealistic will have no effects).

Dataset resolution: 1024*1024, strokes appear at this or higher res.

Recommended weight: 1.0 - 2.0. Go easy on weight, as it might fuck with ADetailer.

Trigger words: ZaUm, texture,

训练模型: NAI，在照片写实模型中没有效果。

图片分辨率: 1024*1024，笔触在高分辨率中出现。

建议权重: 1.0 - 2.0. 当使用ADetailer时，降低权重。

触发词: ZaUm, texture,

Blend in well with what's already inside any given checkpoints. Only downside is that hand gets distorted when far away.

Feedbacks and reviews are appreciated.

Dataset: 极乐迪斯科(风格) | Disco Elysium - Style LoRA - Training Data | Stable Diffusion LoRA | Civitai