V3.1 更新添加人物关键词，liyifeng,yangyang,lixian,wangjunkai,zhangyixing,wulei,Andy Lau,Stephen Chow<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/a1ca94f8-58fc-4a87-9bb6-2ff7852f5bed/width=525/a1ca94f8-58fc-4a87-9bb6-2ff7852f5bed.jpeg" /><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/085de834-5b93-4c9a-b827-96016a4c91bd/width=525/085de834-5b93-4c9a-b827-96016a4c91bd.jpeg" />用了几个新时代的偶像做的男性lora，希望大家喜欢tip1:注意步数再13到28 即可，过低的步数出图会扭曲，过高的步数会出现过拟合，tip2:然后是lora的权重最好再0.4~0.7 之间，太低出图就不像中国人了，太高就很像原图没有随机性，tip3:可以使用低生成-步数，高修复-步数，来生成眼睛相对较小，面部轮廓硬朗的男性最后，记得点赞，点赞，点赞呀，"We created a male LoRa using several new era idols. We hope you like it.Tip 1: Please pay attention to the number of steps between 13 and 28, as too few steps can cause distortion in the output image, and too many steps can lead to overfitting.Tip 2: The weight of LoRa should ideally be between 0.4 and 0.7. If the weight is too low, the output image may not look like a Chinese person, and if it's too high, it may lack randomness and resemble the original image too much.Tip 3: To generate a male image with relatively small eyes and a strong facial contour, you can use a low-generation step count and a high-repair step count.Lastly, please remember to like, like, and like!"：<a target="_blank" rel="ugc" href="http://by.chat">by.chat</a> gpt 3.5补充：对不起，各位，我不知道vae和采样方法的重要性，所以没写，现在补上。、大模型：chilloutmixvae:vae-ft-mse-840000-ema-pruned采样方法：DDIMHere's the translation:Sorry everyone, I didn't mention the importance of VAE and sampling methods earlier, so I'm adding them now.The name of the large model is "chilloutmix", and it uses the VAE model "vae-ft-mse-840000-ema-pruned" with the sampling method "DDIM".

Files with SHA256: 058198c50e007dd518e3d1dabf03d139a8f5986daa859d84ce9d08153b34e963