About Anima v1.0：强度：0.9-1特别注意：不要在不指定画面风格的情况下使用anima base搭配这个lora，你需要使用画师名，或使用风格lora，或使用别的风格稳定的微调模型，否则你可能会得到非常偏向写实风格的图像。还没有和anima的角色lora搭配使用过，还在摸索中。---Strength: 0.9–1Important: Don't use this LoRA with Anima base without specifying an art style. You need to use an artist name, or a style LoRA, or another stable fine-tuned model. Otherwise you may get very realism-leaning images.Haven't tested it with Anima character LoRAs yet—still figuring that out.<hr />About illustrious v1.0：强度：0.8-0.9当与角色lora共同使用时，最好将角色lora的强度降至0.8。碎碎念：<ul><li>1.关于很久没更新新lora：</li></ul>从上次更新到现在，大概已经过了快两个月了。这期间主要是换了新工作，空闲时间变得非常少，工作日只有晚上有一些空闲，但回家又非常非常累，完全不想动......如果继续按照我以前的方法来制作lora，那时间根本不够（主要是预览图太花时间了），所以我趁着618换了新电脑（内存条实在是太贵了，只能买便宜的，但是稍微便宜的内存品控又实在一言难尽，折腾了两周才最终搞定），又花了几天时间更新comfyui工作流。可以预见的后果是：不会被制作预览图折磨了，但以后的lora预览图风格、人物一致性恐怕做不到之前的那么统一。<ul><li>2.关于后续计划：</li></ul>当然是尝试最新最潮的anima......也会重新开始制作z image的lora，之前想着z image后续会推出编辑模型，所以尝试了几个lora之后就没再继续了，但现在看来z的编辑模型是没戏了，那我只能重新开始学习了。新模型后续仍会同步上传到civitai和tensor（civitai现在看起来一副要倒闭的样子，希望它能没事）......---Strength: 0.8-0.9When using this LoRA with character LoRAs, it's best to lower character strength to 0.8.PS:<ul><li>1. Why I haven’t updated new LoRAs for a while</li></ul>It’s been almost two months since my last update. The main reason is that I started a new job, so I have very little free time now. On weekdays, I only have some time at night, but I’m usually super tired when I get home and just don’t feel like doing anything… If I keep making LoRAs the same way as before, there’s simply not enough time (especially because making preview images takes too long). So during the 618 sale, I bought a new computer (RAM was too expensive, so I had to buy cheaper ones, but the quality control for cheaper RAM is really hit or miss — it took me two weeks to finally get it working). I also spent a few days updating my ComfyUI workflow. The likely result: I won’t suffer as much making preview images, but the style, character etc. consistency of future LoRA previews probably won’t be as uniform as before.<ul><li>2. Future plans</li></ul>Of course, I’ll try the latest and trendiest stuff — Anima… I’ll also start making LoRAs for Z-Image again. Earlier, I stopped after a few tries because I thought Z-Image would release an editing model later. But now it seems that editing model isn’t coming, so I’ll have to go back to learning from scratch. I’ll still upload new models to Civitai and Tensor (Civitai looks like it’s about to go under — hope it’ll be okay).<hr />

Files with SHA256: 5e116dc864cfbb139384afe20711a185473392b009adb5dce06f89926b3e4cd3