QQ|Archiver|手机版|小黑屋|AiOMG ( 粤ICP备2022147020号-1 )

GMT+8, 2024-9-21 14:46 , Processed in 0.040158 second(s), 30 queries .

Powered by Discuz! X3.5 Licensed

© 2001-2024 Discuz! Team.

[LoRA] LEOSAM's FilmGirl 胶片风 LoRA--全新VELVIA版本

23
0
1
0
simonlive LV9
正文
发布时间:2024-07-24
模型类型
模型类型: LoRA
基础模型: SD 1.5
最新版本: v3.0
触发词:
下载链接: https://pan.baidu.com/s/1iPZRuzCOffSXPhrgB-AfQA?pwd=qivu
提取码: qivu
源链接: https://www.liblib.art/modelinfo/e44b67fbe1d5c73f7aa22a91adfde9b8?from=search
模型简介:
这是一个用于生成胶片风格AI写真照片的LoRA模型,C站上下载量已近6万,广受大家好评。
该模型与MoonMIX、majicMIX realistic或Chilloumix等写实模型搭配使用,可以生成逼真的胶片风格照片。同时推荐搭配TEXTUAL INVERSION模型:bad_pictures、negative_hand Negative Embedding。这两个Negative Embedding模型不会显著影响到胶片的质感。

2023.7.8 Introducing "The VELVIA 3.0"

更新内容:
Velvia 3.0的主体LoRA模型的训练集图像数量已达到406张,分辨率进一步提升至1024*1024。
本版本主体模型训练使用了Scale weight norms功能,以避免模型过拟合。在另一个备用版本中我也测试了“Prodigy” 优化器, 这个优化器可以在一半的训练总步数下提供与AdamW8bit相当的拟合程度,但可能是我参数设置不合适,训练结果的人脸美感较差,因此最终没有采用。
在高分辨率主体模型基础上,少比例融合了Velvia 2.0模型,同时进行了分层调节,以改善人像肤色。
Update Content:
The number of training images for the main LoRA model of Velvia 3.0 has reached 406, with the resolution further increased to 1024x1024.
In this version, the main model training utilizes "Scale weight norms" functionality to avoid overfitting. I also tested the "Prodigy" Optimizer in another alternative version. Prodigy can provide a fitting degree comparable to AdamW8bit with half the total training steps, but perhaps due to inappropriate parameter settings, the facial aesthetics of the training results were not as good, so it was not adopted in the end.
Based on the high-resolution main model, a small proportion of the Velvia 2.0 model was blended, and hierarchical adjustments were made to improve the skin tone in portraits.
Velvia 2.0的主体LoRA模型的训练集图像分辨率进一步提升至768*1152。
本版本主体模型训练时,加入了15张亚洲年轻男性胶片训练照片,以提高模型在生成男性照片时的真实感。
在高分辨率主体模型基础上,少比例融合了之前两版评价良好的Filmgirl历史版本模型。经过多次测试,这样融合模型的操作可以减少LoRA模型对脸型的影响。

~~~~~~~~~~~~~~~~~~~~~~~o(* ̄︶ ̄*)o~~~~~~~~~~~~~~~~~~~~~~~

示例提示词和参数:
Prompt: <lora:FilmVelvia3:0.6>,ClassicHollywoodglamour,girlinastunningredeveninggown,dazzlingaccessories,(captivatingeyes,redlips,flawlessskin),luxuriousbackdrop,velvetcurtains,dramaticpose,interestingcomposition,spotlightillumination,richshadows,shallowdepthoffield,sparklingbokeh,exquisitedetails,high-resolution,35mmfilm,timelesselegance.portraitphotography,35mmfilm,naturalblurry

Negative prompt: ((worst quality, low quality), bad_pictures, negative_hand-neg:1.2),

Steps: 27, Size: 1024x1664, Seed: 1401714505, Model: majicmixRealistic_v6 , Sampler: DPM++ 2M Karras, CFG scale: 7

本帖子中包含更多资源

您需要 登录 才可以下载或查看,没有帐号?立即注册 微信登录 手机动态码快速登录

x
回复

使用道具

 
您需要登录后才可以回帖 登录 | 立即注册 微信登录 手机动态码快速登录

本版积分规则

快速回复 返回顶部 返回列表