一个厚涂风格的古装lora模型，带点藏色的技巧，色彩比较丰富。还是推荐搭配bjwdd的lora模型使用，bjdww模型在我主页。如果画不出满意的图建议多调整几次权重。这是我用的两组关键词，可以参考一下不搭配bjdww的：1boy, cat ears, male focus, long hair, purple eyes, white hair, sitting, solo, bishounen, looking at viewer, crossed legs, cage, pectorals, birdcage, book, open clothes, holding, &lt;lora:houtu:1&gt;<img src="https://imagecache.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/5971ea4c-68ef-471a-31e3-f9f17d88af00/width=525/5971ea4c-68ef-471a-31e3-f9f17d88af00" />搭配bjdww的：1boy, cat ears, blurry, depth_of_field, blurry_background, long_hair, solo, white_clothes, looking_at_viewer, &lt;lora:houtu:1&gt; &lt;lora:bjdww:0.4&gt;<img src="https://imagecache.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/8732c66d-dbb1-449b-b87c-bde44f137900/width=525/8732c66d-dbb1-449b-b87c-bde44f137900" />1boy, blurry, depth_of_field, blurry_background, long_hair, solo, hanfu, japanese_clothes, looking_at_viewer, &lt;lora:houtu:1&gt; &lt;lora:bjdww:0.5&gt;这个关键词我是直接复制ai反推的，ai把汉服识别成了japanese_clothes，不过实际用的时候，不写japanese_clothes或者写Chinese clothes也都能画出来。<img src="https://imagecache.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/17634405-7cff-4193-8a86-c58385cf7400/width=525/17634405-7cff-4193-8a86-c58385cf7400" />

Files with SHA256: e4d5f07c46a23f920d5741369c7d19475ed0aee4aca2d6af078cfc326ade0f55