Model

v5-NoobAI使用 NoobAI Epsilon-pred 0.5-Version 製作的 Lora，並融合了低權重 v4-rev9，此次訓練相較於之前的版本有更高的穩定度,建議的權重值在 1.0 ,當你使用 NoobAI 系列模型時，權重的...

经典C开头惨败，又又又是个四⭐残次品More weight means more detailToo much detail leads to confusiontry to lower the weight 描述: 训练词语: ...

Hello there! If you're enjoying my project, consider joining to our Ko-fi Memberships. and get ✂️50% discount on Commissions with other benefits. Your support means a lot, and ...

Each version is a unique mix and will give you a different consistent face, they are not building upon the last.These are my Nobodies from SD1.5 retrained to be on SDXL.Please ...

It is a Remix Model of the following amazing models:Analog_diffusionDreamlike_photoreal3DKX_v1.0BAs you see, it strongly enhanced the photorealistic power from Analog and Dream...

木有描述，权重0.6-0.8出图负面词用了EasyNegative请勿用于任何商业用途和其他非法用途，严格遵守各国法律,下载即表示您同意上述协议,点菜群：https://t.me/+0TaXk67vTZ9jNDZl ...

!!! THERE IS A FP16 VERSION IN THE FILES DROPDOWN !!!!!! THE NEWEST VERSION IS v-obsidian-1.0 !!!Some of the images might use embeddings or Loras.If you want more creative resu...

V2做了2个版本有滤镜和无滤镜，区别就是颜色方面,其它差别不大,大家看例图喜欢哪个下哪个就好！触发词：xuer Ancient Chinese sword针对中国剑和剑鞘的强化性训练，写swordsman有机率画出来...

This is a requested character.Your characters may have insufficient training data! Click here.Weight: 1.0Trigger + Appearance: wang liu mei, twin braids, double bun, bun coverD...

■TenganMixV2（Changes from V1）・The output is now higher resolution with changes made to everything except the base model.■TenganMixV2（V1からの変更点）・ベースモデル以外を変...

All the characters from the Gourmet Guild (excluding the MC). You may use weight 0.9 to generate accurate representation of the character.Pecorine (Eustiana von Astraea)Base Tr...

Why did the V6 show up as 3 models? Because I can't make up my mind... REMEMBER ABOUT VAE (I have not baked it into any of the V6 models)V8 - maybe inferior sometimes and train...

Everybody loves striped thigh high socks, right? Unfortunately, I found a great number of the excellent checkpoints on this site have a devil of a time rendering actual stripe...

Euphyllia Magenta from Tensei Oujo to Tensai Reijou no Mahou KakumeiFor best results use:euphyllia_magenta, braid, long silver hairSee also: https://civitai.com/models/12749/an...

如你所见，这个lora可以产生较为干净而且可爱的图像线条比较硬，这使得手或者脚的表现可能有些不如意如果你觉得还算好看的画，就点个赞吧！As you can see, this lora produces a relativel...

Originally posted to HuggingFace by WurzeldiebTextual inversion embedding for SD2.0 and 768-v-ema.ckptConverts even simple prompts into landscape paintingsAll examples are made...

我尝试了新的训练思路,现在服装更加接近原角色,可能要使用0.6或更高的权重I tried new training ideas and now the costume is much closer to the original character ...

One of my first hypernetwork that I'm confident enough to share.Trained on about 40 yandere/gesugao images, focusing on bright eyes and spot colors on grayscale and gradient ba...

你可以通过这个非常cooooooooooooooool的lora生成只属于你自己的带有克味的女孩子，并带有一些艺术气息甚至不是女孩子（bushi一定要增加girl的权重，例如（girl:1.3）,不然你的老婆可能就变...

KeiKomori - Artist StyleThis Lora is trained in the art style of KeiKomori.Some notable games where this artist served as the main illustrator include "ワルキューレロマンツェ ...