Basemodel for 【SDXL】Anime_Pile Driver - v1.0 | Stable Diffusion Checkpoint | Civitai
Training with a 200,000 style image, increased/enhanced the style of about 300 artists, increased/enhanced the concept of 500 characters
quality tag:Best-A,Medium-B,Low-C,Bad-D
The tag training list for the model is in config for the model list
基础模型使用了【SDXL】Anime_Pile Driver - v1.0 | Stable Diffusion Checkpoint | Civitai
使用了20万的风格图片进行训练，增加/增强了约300位画师的风格，增加/增强了500位角色的概念特征
质量标签:Best-A,Medium-B,Low-C,Bad-D
模型的tag训练列表在模型列表的config中
Description
FAQ
Comments (4)
The next major version is currently being trained. The dataset size is 3 million, and the estimated training time is two months.
Wow, it looks promising. What was the focus of the dataset? Variety of styles and knowledge of characters?
Thank you so much! One question please, is stacking multiple quality tags will improve quality?
@trytensor It is enough to use Best-A only in the first part of the positive prompt word. The training amount of the model is not enough to completely eliminate the original quality word
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.