This is a LoRA model for generating film-style AI portrait photos, with over 60,000 downloads on Civitai and widely praised by users.
When used in combination with photorealistic models such as majicMIX realistic, MoonMIX, or Chilloumix, this model can produce realistic film-style photos. It is also recommended to use it in conjunction with TEXTUAL INVERSION models: bad_pictures and negative_hand Negative Embedding. These two Negative Embedding models will not significantly affect the texture of the film.
The currently uploaded Velvia 2.0 version is the latest in the film-style series and offers the most balanced blend of film texture and facial effects. Updates in this version include:
The image resolution of the main LoRA model's training dataset has been further increased to 768x1152.
During the training of the main model, 15 Asian young male film training photos were added to improve the model's realism when generating male photos.
Based on the high-resolution main model, a small proportion of the well-reviewed Filmgirl historical version models from the previous two versions have been blended. After multiple tests, this blended model operation can reduce the LoRA model's impact on facial shapes.
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majicMIX realistic、MoonMIX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。
目前上传的Velvia 2.0版本是胶片风系列模型的最新版本，也是胶片质感与面容效果最为平衡的版本。该版本的更新内容包括：
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。