A style LoRA for the Anima (Cosmos DiT) architecture that refines lighting and mood while preserving strong prompt control.
一个面向 Anima（Cosmos DiT）架构的风格型 LoRA，在强化光影与氛围的同时保持对提示词的强可控性。

Trigger word: vglufeistyle
触发词：vglufeistyle

Recommended base model: Anima-architecture checkpoints
推荐底模：Anima 架构模型

LoRA strength: 0.6 to 1.0 (1.0 recommended; 1.3 may over-expose).
LoRA 强度：0.6 至 1.0（推荐 1.0；1.3 可能过曝）。

Sampling: steps 32, cfg 4, sampler euler, scheduler simple.
采样参数：步数 32、cfg 4、采样器 euler、调度器 simple。

Resolution: 1024 x 1536 (portrait) or similar ratio.
推荐分辨率：1024 x 1536（竖图）或同级比例。

Negative prompt: worst quality, low quality, blurry, extra fingers, bad anatomy.
负向提示词：worst quality, low quality, blurry, extra fingers, bad anatomy。

Usage: place in ComfyUI/models/loras/, load an Anima base model, attach via the Anima LoRA node, and add "vglufeistyle" to your prompt.
使用方法：放入 ComfyUI/models/loras/，加载 Anima 底模，经 Anima LoRA 节点挂载，并在提示词中加入 vglufeistyle。

A style LoRA with no character or outfit binding; stackable with other Anima LoRAs at moderate total weight.
风格型 LoRA，不绑定人物或服装；可与其它 Anima LoRA 叠加，注意总权重适中。

Great for: atmospheric anime illustrations and consistent mood across a set.
适合：富有氛围感的动漫插画，以及统一情绪基调的作品集。

Tips: stable across all four tested strengths (0.3 / 0.6 / 1.0 / 1.3); 0.6-1.0 is the comfortable range.
小贴士：在 0.3 / 0.6 / 1.0 / 1.3 四档强度下均稳定，0.6-1.0 为舒适区间。