TAG:<pre><code>hoshino_ai_\(oshi_no_ko\)
hoshino_ruby
hoshino_aquamarine
kurokawa_akane
memcho</code></pre>v1.0版本使用3000张图片训练。(最新的一张图片来自2024.4.26)为了尽可能缩小图片数量，训练集中不包含多人图片，仅存在单人图片。为了防止角色特征互相污染，使用全部特征标签训练。因为是业余时间在做这个系列，我无法为每一位角色都创建单独的示例图像，因此模型的反馈和优化建议，就拜托大家了。Version 1.0 was trained using3,000 images. (The latest image is from April 26, 2024.)In order to minimize the number of images, the training set does not include images with multiple people. Only single-person images are present.To prevent feature contamination among characters, all feature tags were used for training.Since I'm working on this series in my spare time, I couldn't create separate example images for each character. Therefore, I rely on everyone's feedback and optimization suggestions for the model.Please note that this model's content includes some covers that have been rated as PG-13/R/XPlease choose the appropriate age rating based on your needs.Training model: <a target="_blank" rel="ugc" href="https://civarchive.com/models/257749?modelVersionId=290640">Pony</a>Recommend model: Pony Diffusion V6 XL/T-ponynai3/AutismMix SDXLTraining resolution: 1024*1024Recommended output resolution: 1024*1280特别鸣谢/算力赞助 : 墩墩的4090Discord : <a target="_blank" rel="ugc" href="https://discord.com/invite/CFpyHceF52">https://discord.com/invite/CFpyHceF52</a>Recommended dataset collection and processing tools: <a target="_blank" rel="ugc" href="https://github.com/deepghs/waifuc">https://github.com/deepghs/waifuc</a>Possibly the most user-friendly dataset processing UI currently in use： <a target="_blank" rel="ugc" href="https://github.com/Eugeoter/sd-dataset-">https://github.com/Eugeoter/sd-dataset-</a>各个角色字符在训练集中的分布：The distribution of characters in the training set is as follows:<pre><code> hoshino_ai_\(oshi_no_ko\): 1041
 arima_kana: 1064
 kurokawa_akane: 285
 hoshino_ruby: 253
 memcho: 109
 hoshino_aquamarine: 78</code></pre>

Files with SHA256: d9d7f330f0543f31177ad69fbc6f8dd5026547d23279df9f7f6b57a06f22a41a