起初我训练了一个关于 “高桥留美子” 笔下角色的风格 Lora，但因为效果不是很好，最后将其与我另一个 Lora <a rel="ugc" href="https://civarchive.com/models/642325?modelVersionId=718459">时髦的独特的(stylish&amp;chic)</a><a rel="ugc" href="https://civarchive.com/models/669678/eye-trail"> </a>混合，从而得到了这个 Lora。该 Lora 会倾向生成扁平的平涂风格，拥有高对比高亮度的效果。同时生成的人物，非常接近 “犬夜叉” 与 “乱马1/2” 中的角色外貌。根据使用的大模型，生成效果也有所不同，部分自带风格强烈的大模型可能效果不明显。请多多尝试。<hr />Initially, I trained a LoRA focused on the style of characters from Rumiko Takahashi's works, but since the results weren’t ideal, I eventually mixed it with another one of my LoRAs : <a target="_blank" rel="ugc" href="https://civarchive.com/models/642325?modelVersionId=718459">时髦的独特的(stylish&amp;chic)</a><a target="_blank" rel="ugc" href="https://civarchive.com/models/669678/eye-trail"> </a>. This blend resulted in the current LoRA.This LoRA tends to generate a cel-shading style with high contrast and brightness. The characters it produces closely resemble those from Inuyasha and Ranma 1/2.The output may vary depending on the base model used, and it might not be as effective with models that have a strong pre-existing style. Please feel free to experiment with different settings.

Files with SHA256: 09710d63e2e79153b3728115e2f97de7ef355e34efdb3917d758d264aabf2289