在小红车找到的超色二游风格，鉴于Anima模型还没有能让我满意的油丝质感，特化丝袜来训练了这个lora（感谢画师们的绝色美图）

有什么问题可以反馈给咱，线上训练反反复复干了400多，大家喜欢的话来点图图和赞，给咱回回血）

This lora's training cost me a lot QAQ.

I really need your like！

触发词：@crosscore/glossy stockings（油亮丝袜质感）/see through stocking（透丝）

推荐权重：0.6-1.0

这个lora的数据集脸比较小，低一点权重和其他lora搭配效果更好，0.6的情况下丝袜质感依然突出，画风影响相对减少

使用方法：

- 推荐正向前缀：masterpiece, best quality, score_9, score_8, highres, absurdres, @crosscore, ...

- 推荐负向：worst quality, low quality, score_1, score_2, score_3, artist name, blurry, jpeg artifacts

- 推荐权重：1.0（可按喜好调 0.8-1.3）

- 采样：30 步、CFG 4、Euler（simple 调度）、1024x1536、ModelSamplingAuraFlow shift 1-3

- 任意角色可用，画风可迁移到训练集之外的角色

Title: crosscore style - 2D Game CG / Glossy Stockings LoRA for Anima

A style LoRA trained on Anima Base v1.0. Reproduces a clean 2D game-CG look

with glossy, see-through stockings and oil-slick highlights.

Trigger word: @crosscore

Usage:

- Suggested positive prefix: masterpiece, best quality, score_9, score_8, highres, absurdres, @crosscore, ...

- Suggested negative: worst quality, low quality, score_1, score_2, score_3, artist name, blurry, jpeg artifacts

- Recommended weight: 1.0 (adjust 0.6-1.0 to taste)

- Sampler: 30 steps, CFG 4, Euler (simple scheduler), 1024x1536, ModelSamplingAuraFlow shift 1-3

- Works on any character; style also transfers to non-training characters

Training details:

- Base model: Anima Base v1.0

- Network: LoRA (networks.lora_anima), dim 64 / alpha 64, UNet-only

- Dataset: 158 images x 3 repeats (474 per epoch), resolution 1536x1536, bucket 1024-1536

- Optimizer: AdamW (weight decay 0.01), LR 2e-5, constant scheduler, warmup 0

- Timestep: logit_normal, weighting logit_normal, discrete_flow_shift 3.0

- Loss: L2, gradient checkpointing, bf16, cache latents

- 40 epochs