<h2 id="euler-a-6102n4f3k">推荐采样器：Euler a</h2><h2 id="sampler-is-recommendedeuler-a-r3cwwax67">Sampler is recommended：Euler a</h2>Steps:30CFG:8放大1.5，4x_NMKD-Superscale-SP_178000_G<h3 id="-voy3kah3p">推荐正反提示词</h3><h3 id="positive-and-negative-prompts-are-recommended-nqq1vps14">Positive and negative prompts are recommended</h3><pre><code>score_9,score_8_up,score_7_up,</code></pre><pre><code>score_6,score_5,score_4,</code></pre><ul><li>V8.1：由于8.0的某项权重过高导致人物不是很漂亮，修改了权重大大改观</li><li>V8.1: Due to the high weight of a certain item in 8.0, the character is not very beautiful, and the weight has been greatly changed</li><li>V8.0：黑暗</li><li>V8.0: Darkness</li><li>V7.0：颜色没有那么鲜艳了加入了新的Lora</li><li>V7.0: The colors are not so bright anymore, and the new Lora has been added</li><li>V6.6：将4,5,6版本融合加入新训练的Lora。</li><li>V6.6: Fusion versions 4, 5, and 6 to the newly trained Lora.</li><li>V6.0：增强细节，光线偏暗</li><li>V6.0: Enhanced detail, dark light</li><li>V5.0：制作了一个人物手部不容易出错的模型NiceBody模型画风一般，然后结合了3.0模型的一些画风效果还不错。</li><li>V5.0: Made a model of a character whose hands are not easy to make mistakes, and the NiceBody model has a general painting style, and then combined with some of the 3.0 model's painting style, the effect is not bad.</li><li>V4.0：风格变化非常大，加入了自己新训练的LORA</li><li>V4.0: The style has changed very much, and I have added my own newly trained LORA</li><li>V3.1：由于3.0问题有点多所以改进了3.0并且融入了新训练的Lora可以出有意思的图，风格不一样了。</li><li>V3.1: Because 3.0 has a lot of problems, Lora has improved 3.0 and incorporated new training to produce interesting diagrams,The style is different.</li><li>V3.0：开始传错了模型，把中间模型传上来了，换好了，增强了细节。</li><li>V3.0: The wrong model was transmitted at the beginning, and the intermediate model was uploaded, replaced it, and the details were enhanced.</li><li>V2.6：不那么真实但是可以产出有意思的图</li><li>V2.6: Less realistic but can produce interesting graphs</li></ul><ul><li>V2.5：全面提升真实度</li><li>V2.5: Comprehensively improve realism</li><li>V2.1：手部不会那么崩了。</li><li>V2.1: Hands don't break so much.</li><li>V2.0：重新融合了模型，化繁为简，效果很不错。</li><li>V2.0: Re-integrated the model, simplifying the complex, and the effect is very good.</li><li>V1.7：细节进一步增强</li><li>V1.7: Further enhancements to details</li><li>V1.6：1.5和1.0的结合，皮肤白皙，细节增强。</li><li>V1.6: A combination of 1.5 and 1.0 with fair skin and enhanced details.</li><li>V1.5：修复了一些手部问题和出现多只脚的几率，增强了细节。</li><li>v1.5: Fixed some hand issues and the chance of multiple feet, enhanced details.</li></ul>AI新手学习了几个月，自己训练了一些LORA，融合了几个模型，感觉不错分享给各位，有些模型融合了效果很好，但是那些模型不允许融合分享，只能分享这个了，这个模型加入自己的Lora后人物的皮肤更加白皙，更加趋向东方人面孔，手脚也更加不容易出错，用了四个模型相互融合，感谢这四个模型的创造者。English is not good, the following is translated by translation software, and I don't know if there will be ambiguity.AI novices have been learning for a few months, trained some LORA by themselves, fused a few models, and it feels good to share with you, some models are fused and work very well, but those models are not allowed to be fused and shared, so you can only share this, and the skin of the characters after this model joins their own Lora is fairer, more inclined to oriental faces, and their hands and feet are less prone to mistakes,It took four models to blend together, thanks to the creators of these four models.<h1 id="duchaiten-pony-xl3.0-tq0gzmeoy"><a target="_blank" rel="ugc" href="https://civarchive.com/models/376450?modelVersionId=481945">DucHaiten-Pony-XL的3.0版本</a></h1><h1 id="pianomixcosplay1.0-1246m1rfn"><a target="_blank" rel="ugc" href="https://civarchive.com/models/400801?modelVersionId=492976">PianoMix的cosplay1.0版本</a></h1><h1 id="nova-reality2.0-jytnsmafb"><a target="_blank" rel="ugc" href="https://civarchive.com/models/453428/nova-reality?modelVersionId=573349">Nova Reality2.0</a></h1><h1 id="cyberrealistic-pony4.0-2ysq181zh"><a target="_blank" rel="ugc" href="https://civarchive.com/models/443821/cyberrealistic-pony?modelVersionId=595775">CyberRealistic Pony4.0</a></h1><h2 id="v1.7-1a5kzadfr">v1.7新融合了两个模型感谢两位作者</h2><h1 id="satpony-i39ahgbcv"><a target="_blank" rel="ugc" href="https://civarchive.com/models/527782/satpony?modelVersionId=617749">SatPony</a></h1><h1 id="ponykaine_paint-tqosiyxzo"><a target="_blank" rel="ugc" href="https://civarchive.com/models/494472/ponykainepaint?modelVersionId=614375">PonyKaine_Paint</a></h1><h2 id="v2.0-re0w0hqdu">v2.0用了新的模型感谢作者</h2><h1 id="duchaiten-pony-real-ch51dlg7u"><a target="_blank" rel="ugc" href="https://civarchive.com/models/477851/duchaiten-pony-real">DucHaiten-Pony-Real</a></h1><h2 id="v2.6-ip6ffes78">v2.6添加了新的模型感谢作者</h2><h1 id="satpony-niji_v2-cyc1crevy"><a target="_blank" rel="ugc" href="https://civarchive.com/models/582502/satpony-niji?modelVersionId=667513">SatPony-Niji_V2</a></h1><h2 id="v3.0-86ujyrxon">v3.0添加了新的模型感谢作者</h2><h1 id="satpony-journey2.0-hcm9n8m0h"><a target="_blank" rel="ugc" href="https://civarchive.com/models/595554/satpony-journey?modelVersionId=706765">SatPony-Journey2.0</a></h1>

3010nc-xx-MixPony

3010ncXxMixpony_v20VAE.safetensors

v2.0+VAE

v8.1

v8.0

v7.0

v6.6

v6.0

v5.0

v4.0

v3.1

v3.0

v2.6

v2.5

v2.1

v2.0

v1.7

v1.6

v1.5

v1.0

不用正反提示词也能出不错的图而且非常听话。------ You don't need Positive Prompts and Negative Prompt to make good pictures and very obedient.

Files with SHA256: 166882ac8a6b78ccb51835d2027585822b1565623b1fa7377c1b624f8f5b5fae