LoRA de usuario

Community DiT
Personaje

[VRChat] Karin | カリン | あまとうさぎ

#VRChat
[VRChat] Karin | カリン | あまとうさぎ - AI Model cover image
Subido en
20 sept 2026, 16:35
Utiliza
186
Reseñas
Sin reseñas
Palabras clave

karin \(vrchat\)

Permisos
  • Permitir la generación y el intercambio de imágenes
  • Permitir que los usuarios descarguen su modelo
  • Usos comerciales

Descripción

IntroductionThis LoRA is trained on Karin (カリン), a VRChat avatar created by Amatousagi (あまとうさぎ).UsageRecommended LoRA strength: 1.0Trigger word: karin (vrchat)In ComfyUI, escape the parentheses:karin (vrchat)Default outfit:white sweatshirt, long sleeves, pink sleeve bows, black suspenders, black pleated skirt, white sneakers, fox tailTraining DetailsTrainer: sorryhyun/anima_lora v1.17.1Dataset: 51 captioned images, with 23 higher-quality images sampled twice and 28 images sampled onceLoRA rank / alpha: 32 / 128Learning rate: 2e-5Optimizer: AdamWScheduler: Cosine, with 5% warmupBatch size: 1Released checkpoint: Step 1,392Precision: BF16Resolution: Aspect-ratio buckets at approximately 1 megapixelGPU: NVIDIA RTX 3090简介这是基于 Anima Base 1.0 训练的 Karin(カリン)角色 LoRA。Karin 是由 あまとうさぎ 制作的 VRChat 虚拟形象。使用推荐 LoRA 权重: 1.0激活词: karin (vrchat)在 ComfyUI 中需要转义括号:karin (vrchat)常服:white sweatshirt, long sleeves, pink sleeve bows, black suspenders, black pleated skirt, white sneakers, fox tail训练参数训练工具: sorryhyun/anima_lora v1.17.1数据集: 51 张带 caption 的图片,其中 23 张高质量图片重复采样 2 次,其余 28 张采样 1 次LoRA rank / alpha: 32 / 128学习率: 2e-5优化器: AdamW学习率调度: Cosine,5% warmupBatch size: 1发布权重: 第 1,392 步训练精度: BF16训练分辨率: 约 1 M 像素的宽高比分桶显卡: NVIDIA RTX 3090

Comentarios