🐰 BUNNY's LoRA Series
✦ ABOUT THIS KREA2 MODEL / 关于KREA2模型 ✦
This model (trained on KREA2) was trained using Civitai's on-site trainer.
这个基于 KREA2 的模型是在 C 站云端完成训练的。
However, the core dataset was meticulously cleaned, curated, and tagged using my personal workflow.
不过，最核心的数据集完全是由我通过独家流程精细清洗、筛选与打标完成的。
This ensures maximum fidelity, clean aesthetics, and excellent prompt responsiveness.
这保证了模型具备最高的神韵还原度、干净的审美风格以及极佳的提示词响应度。
✦ SUPPORT FUTURE TRAININGS (NEED YOUR BUZZ!) / 助力后续训练（求 Buzz 投喂！）✦
Training on-site requires a continuous stream of platform compute!
在 C 站云端进行训练需要源源不断的平台算力支持！
If you enjoy this LoRA, please consider dropping some Buzz or leaving a 5-star review.
如果你喜欢这个 LoRA，请考虑投喂一些 Buzz 或留下一个五星好评。
Every bit of Buzz directly fuels my cloud training budget for faster updates and higher quality.
你们投喂的每一份 Buzz 都会直接转化为我的云端算力，带来更快的更新与更高质量的模型。
Hope you like it, and feel free to share your showcase renders below!
希望你们喜欢，非常欢迎在下方晒出你们的跑图作品！
推荐权重 (Weight):
ANIMA BASE/Aesthetic : 0.85-0.95 CFG:3.5-4.5
Turbo : 0.75-0.9 + High Resolution CFG:1
When using it on Base, pair it with the POV Lora to adjust the camera distance.
recommend using 3D Game Vision LORA to enhance 3D graphics performance.
🔑 触发词 (Trigger Words):wom1n
面部及发型(Face & Hair1):
wom1n,blonde hair, short hair, bob cut, messy hair, wavy hair,面部及发型(Face & Hair2):
wom1n,blonde hair, short hair, bob cut, wavy hair, hoop earrings, red lips,FBI服装 (FBI Outfit):
glasses, black blazer, white shirt, id card, black pants, 服装2 (Outfit2):
white off-shoulder blouse, black miniskirt, belt, black elbow gloves,I’d love to see what you create! Please feel free to share your generations in the review section. Your feedback and ratings inspire me to create more!
If you enjoy my works and want to support my future creations, any tip is greatly appreciated! https://ko-fi.com/fourbunny☕💛
Description
✦ ABOUT THIS MODEL / 关于本模型 ✦
This model (trained on KREA2) was trained using Civitai's on-site trainer.
这个基于 KREA2 的模型是在 C 站云端完成训练的。
However, the core dataset was meticulously cleaned, curated, and tagged using my personal workflow.
不过，最核心的数据集完全是由我通过独家流程精细清洗、筛选与打标完成的。
This ensures maximum fidelity, clean aesthetics, and excellent prompt responsiveness.
这保证了模型具备最高的神韵还原度、干净的审美风格以及极佳的提示词响应度。
✦ SUPPORT FUTURE TRAININGS (NEED YOUR BUZZ!) / 助力后续训练（求 Buzz 投喂！）✦
Training on-site requires a continuous stream of platform compute!
在 C 站云端进行训练需要源源不断的平台算力支持！
If you enjoy this LoRA, please consider dropping some Buzz or leaving a review.
如果你喜欢这个 LoRA，请考虑投喂一些 Buzz 或留下一个好评。
Every bit of Buzz directly fuels my cloud training budget for faster updates and higher quality.
你们投喂的每一份 Buzz 都会直接转化为我的云端算力，带来更快的更新与更高质量的模型。
Hope you like it, and feel free to share your showcase renders below!
希望你们喜欢，非常欢迎在下方晒出你们的跑图作品！