KBTIT Lora
файл на civitaiстраница на civitaiNone
None
SHA256: FACAEF321EDAB897F7D7C6DFCE90C604999B29C2E7BCB1AC3DFCF242AF7226B0
A really fun model to play with, it can get really creative sometimes!
Optimal values:0.5-0.8
Enjoy!
If you like it, you are free to share your results, I like seeing everyone's results!:)
A really fun model to play with, it can get really creative sometimes!
Optimal values:0.5-0.8
Enjoy!
If you like it, you are free to share your results, I like seeing everyone's results!:)
straddlingfrench kisstongue kissgirl on top
We comply with KIZUNA AI secondary creation guidelines.
キズナアイ キャラクター利用規約に準拠します。
https://kizunaai.com/download/guideline/
Please read the terms carefully.
規約を一読するようお願い致します。We comply with KIZUNA AI secondary creation guidelines.
キズナアイ キャラクター利用規約に準拠します。
https://kizunaai.com/download/guideline/
Please read the terms carefully.
規約を一読するようお願い致します。kizuna ai
V2 removed many images with messed lines from the data set and changed the batch size for better results.
V2 sample image:
V1 Sample images:
V2 removed many images with messed lines from the data set and changed the batch size for better results.
V2 sample image:
V1 Sample images:
Gain the ability to trace on like the best drunken Master out there! I actually messed up on the training data set of this Lora... not really sure to what extent but it seems to be mainly focused on swords. I liked the results it generated so still wanted to share. Testing a normal one... but that has its own problems.
Trained on random Noble Phantasms. I'd start with a weight of .80. Activation word is dmarble but you can try without it. Requires gacha. As the great Shirou Emiya said, fake it till you make it. Hope you like it!
Example Prompt:
<lora:ldmarble-22:0.8>a detailed sword, dmarble, intricate design, weapon, no humans, sunlight, scenery, light rays, fantasy, sharp focus, extreme details
<lora:ldmarble-22:0.8>a detailed sword, no humans, cloud, blue sky, cloudy sky, scenery, light particles, fantasy, rainbow
<lora:ldmarble-22:0.8>a detailed weapon sword, sunlight, cloudy sky, grass, nature, scenery, light rays, rock, mountain, sunbeam, waterfall, cliff
Neg:
(blurry:1.3), monochrome, extra limbs, distorted, discolored, deformed, disfigured, watermark, (bad hands), (Worst quality, low quality:1.4)
Gain the ability to trace on like the best drunken Master out there! I actually messed up on the training data set of this Lora... not really sure to what extent but it seems to be mainly focused on swords. I liked the results it generated so still wanted to share. Testing a normal one... but that has its own problems.
Trained on random Noble Phantasms. I'd start with a weight of .80. Activation word is dmarble but you can try without it. Requires gacha. As the great Shirou Emiya said, fake it till you make it. Hope you like it!
Example Prompt:
<lora:ldmarble-22:0.8>a detailed sword, dmarble, intricate design, weapon, no humans, sunlight, scenery, light rays, fantasy, sharp focus, extreme details
<lora:ldmarble-22:0.8>a detailed sword, no humans, cloud, blue sky, cloudy sky, scenery, light particles, fantasy, rainbow
<lora:ldmarble-22:0.8>a detailed weapon sword, sunlight, cloudy sky, grass, nature, scenery, light rays, rock, mountain, sunbeam, waterfall, cliff
Neg:
(blurry:1.3), monochrome, extra limbs, distorted, discolored, deformed, disfigured, watermark, (bad hands), (Worst quality, low quality:1.4)
dmarble
Introduced in 1998 at the Geneva Motor Show, the E39 generation of the M5 was the first M5 to use a V8 engine, resulting in an increase in power output to 294 kW (400 PS).[37] It is also the first M5 to use aluminium front suspension components and a multi-link rear suspension. Production began in October 1998.
Rear 3/4 view
BMW S62 V8 engine
BMW M5 E39 Interior
Production totalled to 20,482 cars from 1999 to 2003. Unlike its predecessors, the M5 was produced on the same assembly line as the regular 5 Series models at the Dingolfing factory in Germany.
The official performance figures are 0–97 km/h (60 mph) acceleration time of 4.8 seconds and an electronically limited top speed of 250 km/h (155 mph).[38][39] In testing, an unrestricted M5 reached a top speed in excess of 300 km/h (186 mph).[40] The E39 M5 recorded a Nürburgring lap time of 8:20.[38]
The M5 received the September 2000 facelift (for the 2001 model year) at the same time as the standard E39 models.[41] Changes included halogen "corona rings" in headlights (often called "Angel Eyes"), LED tail-lights and various interior upgrades. The mechanical specification was unchanged. For the subsequent two model years, changes were limited to the addition of new exterior colours (from September 2001) and the upgrade to a DVD-based navigation system (from September 2002).[41]
Production of a "Touring" (wagon/estate) E39 M5 model was evaluated by BMW, and at least one prototype was developed (in Titanium Silver with a Black Exclusive leather interior). However the Touring did not reach production, due to financial considerations.[42]
Introduced in 1998 at the Geneva Motor Show, the E39 generation of the M5 was the first M5 to use a V8 engine, resulting in an increase in power output to 294 kW (400 PS).[37] It is also the first M5 to use aluminium front suspension components and a multi-link rear suspension. Production began in October 1998.
Rear 3/4 view
BMW S62 V8 engine
BMW M5 E39 Interior
Production totalled to 20,482 cars from 1999 to 2003. Unlike its predecessors, the M5 was produced on the same assembly line as the regular 5 Series models at the Dingolfing factory in Germany.
The official performance figures are 0–97 km/h (60 mph) acceleration time of 4.8 seconds and an electronically limited top speed of 250 km/h (155 mph).[38][39] In testing, an unrestricted M5 reached a top speed in excess of 300 km/h (186 mph).[40] The E39 M5 recorded a Nürburgring lap time of 8:20.[38]
The M5 received the September 2000 facelift (for the 2001 model year) at the same time as the standard E39 models.[41] Changes included halogen "corona rings" in headlights (often called "Angel Eyes"), LED tail-lights and various interior upgrades. The mechanical specification was unchanged. For the subsequent two model years, changes were limited to the addition of new exterior colours (from September 2001) and the upgrade to a DVD-based navigation system (from September 2002).[41]
Production of a "Touring" (wagon/estate) E39 M5 model was evaluated by BMW, and at least one prototype was developed (in Titanium Silver with a Black Exclusive leather interior). However the Touring did not reach production, due to financial considerations.[42]
M52002
This is not a necessary, but If you like my Lora, please consider to donate or even Comissioning your own requested Lora on My Ko-fi
https://ko-fi.com/warrensky/commissions
Please give 5 star reviews by sharing your picture creation if you like it,
Use any checkpoint based on SD1.5
its 768,768 image
turn on Hires.fix (or not, please experiment with it yourself)
Hires steps: 14-17, denoising:04-65 ish.. just try it yourself
CGE: 5-7
Use DPM++ 2M Karras or DPM++ SDE Karras
Use 20-ish sampling steps
Strength: 0.6-1
I suggest turn off the restore faces check mark for better results of this Lora
She might look like someone from real life, but she is not what you think. She is a totally fictional character.
It is prohibited to use this model for commercial purposes and any scenarios of illegal acts and purposes.
本人原则上不支持使用此模型生成任何成人内容。用户在使用本模型时，请务必遵守当地法律法规，勿侵犯他人的名誉权、隐私权或肖像权。
I generally do not support the use of this model to generate any adult content. When using this model, users should abide by local laws and regulations and shall not violate any other's right to reputation, privacy or portrait.
私は一般的に、このモデルを使用して成人コンテンツを生成することをサポートしません。 このモデルを使用する場合は、ユーザーは当地の法律や規制を守り、他人の名誉権、プライバシー権または肖像権を侵害しないようにしなければなりません。
This is not a necessary, but If you like my Lora, please consider to donate or even Comissioning your own requested Lora on My Ko-fi
https://ko-fi.com/warrensky/commissions
Please give 5 star reviews by sharing your picture creation if you like it,
Use any checkpoint based on SD1.5
its 768,768 image
turn on Hires.fix (or not, please experiment with it yourself)
Hires steps: 14-17, denoising:04-65 ish.. just try it yourself
CGE: 5-7
Use DPM++ 2M Karras or DPM++ SDE Karras
Use 20-ish sampling steps
Strength: 0.6-1
I suggest turn off the restore faces check mark for better results of this Lora
She might look like someone from real life, but she is not what you think. She is a totally fictional character.
It is prohibited to use this model for commercial purposes and any scenarios of illegal acts and purposes.
本人原则上不支持使用此模型生成任何成人内容。用户在使用本模型时，请务必遵守当地法律法规，勿侵犯他人的名誉权、隐私权或肖像权。
I generally do not support the use of this model to generate any adult content. When using this model, users should abide by local laws and regulations and shall not violate any other's right to reputation, privacy or portrait.
私は一般的に、このモデルを使用して成人コンテンツを生成することをサポートしません。 このモデルを使用する場合は、ユーザーは当地の法律や規制を守り、他人の名誉権、プライバシー権または肖像権を侵害しないようにしなければなりません。
此模型训练的是上半身的，因此推荐只生成上半身的图像，全身图片效果很大可能不会太好？建议以这些TAG为基础进行扩展：masterpiece, best quality,beautifly,1girl,cinematic angle, upper body,
权重的话建议高一点0.6以上。
负面TAG建议有：long neck，因为暂时还没有解决原训练集有点长脖子的问题
This model trains the upper body, so it is recommended to generate only the upper body image, while the whole body image has a great effect, which may not be good??
It is suggested to extend the model based on these tags: Masterpiece, best quality, the beautifly, 1 girl, cinematic Angle, upper body,
higher weights that if more than 0.6
(Negative prompt: long neck),Because we haven't solved the long neck problem in the original training set yet
此模型训练的是上半身的，因此推荐只生成上半身的图像，全身图片效果很大可能不会太好？建议以这些TAG为基础进行扩展：masterpiece, best quality,beautifly,1girl,cinematic angle, upper body,
权重的话建议高一点0.6以上。
负面TAG建议有：long neck，因为暂时还没有解决原训练集有点长脖子的问题
This model trains the upper body, so it is recommended to generate only the upper body image, while the whole body image has a great effect, which may not be good??
It is suggested to extend the model based on these tags: Masterpiece, best quality, the beautifly, 1 girl, cinematic Angle, upper body,
higher weights that if more than 0.6
(Negative prompt: long neck),Because we haven't solved the long neck problem in the original training set yet
Weight: around 0.7-1.0
Main costume: pink hair, buns, blue jacket, green skirt
School uniform: pink hair, buns, white shirt, sweater vest, green skirt
Works fine with both anime and realistic (tried on deliberate2, images awaits moderation)
Tags are not very neat, mostly automatic ones are used, i'l retrain at 768x768 in a week or two.
Happy sugar life - The series is centered on high school girl Satō Matsuzaka (Satou Matsuzaka) who befriends a mysterious little girl named Shio, who was abandoned by her mother, and immediately becomes highly attached to her. They both agree to live together in Satō's apartment, and the former vows to protect that feeling of love, even if it means committing crimes or even killing people.
Weight: around 0.7-1.0
Main costume: pink hair, buns, blue jacket, green skirt
School uniform: pink hair, buns, white shirt, sweater vest, green skirt
Works fine with both anime and realistic (tried on deliberate2, images awaits moderation)
Tags are not very neat, mostly automatic ones are used, i'l retrain at 768x768 in a week or two.
Happy sugar life - The series is centered on high school girl Satō Matsuzaka (Satou Matsuzaka) who befriends a mysterious little girl named Shio, who was abandoned by her mother, and immediately becomes highly attached to her. They both agree to live together in Satō's apartment, and the former vows to protect that feeling of love, even if it means committing crimes or even killing people.
satoubuns, double bunsblue jacketwhite shirt, sweater vestglowing eyes
V3更新：
基于anythingV5底模制作，例图使用mix-pro4直出，大幅度改善了过拟合、换装难和娘化的问题
使用方法：
因为魈的角色特征较为复杂，还原原版服装，推荐权重0.8~0.9，触发词为：1boy, male focus, yellow eyes, (official:1.2)
换装则推荐较低权重0.6~0.8，并拉高服装权重，触发词为：1boy, male focus, yellow eyes
translate by GPT：
V3 Update:
Based on the anythingV5 base model, the example images are output using mix-pro4, which greatly improves the problems of overfitting, difficulty in changing costumes, and feminization.
How to use:
Due to the complex character features of Xiao, it is recommended to use a weight of 0.8-0.9 to restore the original costume, with trigger words：1boy, male focus, yellow eyes, (official:1.2).
For changing costumes, it is recommended to use a lower weight of 0.6-0.8 and increase the weight of the clothing, with trigger words：1boy, male focus, yellow eyes
V3更新：
基于anythingV5底模制作，例图使用mix-pro4直出，大幅度改善了过拟合、换装难和娘化的问题
使用方法：
因为魈的角色特征较为复杂，还原原版服装，推荐权重0.8~0.9，触发词为：1boy, male focus, yellow eyes, (official:1.2)
换装则推荐较低权重0.6~0.8，并拉高服装权重，触发词为：1boy, male focus, yellow eyes
translate by GPT：
V3 Update:
Based on the anythingV5 base model, the example images are output using mix-pro4, which greatly improves the problems of overfitting, difficulty in changing costumes, and feminization.
How to use:
Due to the complex character features of Xiao, it is recommended to use a weight of 0.8-0.9 to restore the original costume, with trigger words：1boy, male focus, yellow eyes, (official:1.2).
For changing costumes, it is recommended to use a lower weight of 0.6-0.8 and increase the weight of the clothing, with trigger words：1boy, male focus, yellow eyes
1boy, male focus, yellow eyes,official
Basically I tried to make a lora to imitate the style of mizuryu kei, it's the first lora I've done and I hope you like it.
I'll try to improve it as soon as I know how the training works
Basically I tried to make a lora to imitate the style of mizuryu kei, it's the first lora I've done and I hope you like it.
I'll try to improve it as soon as I know how the training works
mizuryuslut
Anime lora trained on NAI.
You can choose the color of the bikini like red bikini, black bikini, white bikini.
Anime lora trained on NAI.
You can choose the color of the bikini like red bikini, black bikini, white bikini.
mouth holdbikini
LoRA of Murasaki Shikibu in FGO, works between 0.7-1.0
FGO紫式部的模型，推荐权重0.7-1.0
-
Supports 8 sets of Official Outfits in total, 3 sets for Caster class + 3 sets for Rider class + 2 sets of Craft Essences.
支持共计8套官方外观，C阶3套+R阶3套+礼装外观2套
-
Caster1=Caster1, black long dress
Caster2=Caster2, Plaid pattern coat, black and white coat
Caster3=Caster3, junihitoe kimono, straight hair, hime cut
Rider1=Rider1, Murasaki Swimsuit, single thighhigh, short hair
Rider2=Rider2, JK Uniform, glasses, pink jacket, eyepatch, bandage
Rider3=Rider3, aqua bikini, striped_thigh-highs, gloves, skirt, white sleeves
Purple Eye-紫眸=PurpleEye, Murasaki bikini
Thank You for Your Hard Work!-辛苦啦！= GoodJob, Murasaki Office Lady, straight hair, glasses
-
Feel free to post any idea or character model request in the Discussion
有模型的需求可以B站私信或者在这里留言给我
id：不吃青蛙的阿龙
有空就会做
LoRA of Murasaki Shikibu in FGO, works between 0.7-1.0
FGO紫式部的模型，推荐权重0.7-1.0
-
Supports 8 sets of Official Outfits in total, 3 sets for Caster class + 3 sets for Rider class + 2 sets of Craft Essences.
支持共计8套官方外观，C阶3套+R阶3套+礼装外观2套
-
Caster1=Caster1, black long dress
Caster2=Caster2, Plaid pattern coat, black and white coat
Caster3=Caster3, junihitoe kimono, straight hair, hime cut
Rider1=Rider1, Murasaki Swimsuit, single thighhigh, short hair
Rider2=Rider2, JK Uniform, glasses, pink jacket, eyepatch, bandage
Rider3=Rider3, aqua bikini, striped_thigh-highs, gloves, skirt, white sleeves
Purple Eye-紫眸=PurpleEye, Murasaki bikini
Thank You for Your Hard Work!-辛苦啦！= GoodJob, Murasaki Office Lady, straight hair, glasses
-
Feel free to post any idea or character model request in the Discussion
有模型的需求可以B站私信或者在这里留言给我
id：不吃青蛙的阿龙
有空就会做
Caster1, black long dressCaster2, Plaid pattern coat, black and white coatCaster3, junihitoe kimono, straight hair, hime cutRider1, Murasaki Swimsuit, single thighhigh, short hairRider2, JK Uniform, glasses, pink jacket, eyepatch, bandage,Rider3, aqua bikini, striped_thigh-highs, gloves, skirt, white sleevesPurpleEye, Murasaki bikiniGoodJob, Murasaki Office Lady, straight hair, glasses
This is LoRA, in which I try to depict one of the most popular Bleach characters. Activation tag is "rangiku".
Please, write all pros and cons of the LoRA, your comments and reviews allow me to remake the LoRA (cause I have no possibility to check it properly) and to make my future loras better.
This is LoRA, in which I try to depict one of the most popular Bleach characters. Activation tag is "rangiku".
Please, write all pros and cons of the LoRA, your comments and reviews allow me to remake the LoRA (cause I have no possibility to check it properly) and to make my future loras better.
rangiku
myfriend face,share everyone
myfriend face,share everyone
womanlora
Atelier Escha & Logy - Escha エスカ メーリエ LoRA
Escha character LoRA. Wish you like this LoRA. Trained by samples from official artwork and animation.
Recommended CFG: 6.6-6.5
Recommended weights: 0.8-1
Recommended prompt: choker, hat, short hair, twintails,
Honestly, I would have liked to add more detail to the character, but I had a hard time finding a more detailed costume design in the animation. And, because Escha has a tail-like decoration, it was difficult to generate while training and drawing. I will try to solve this problem in a later version by removing some of the tags.
NOTE: The first version is good for generating avatar/portrait/upper body, possibly, there may be a problem in producing full body images
Atelier Escha & Logy - Escha エスカ メーリエ LoRA
Escha character LoRA. Wish you like this LoRA. Trained by samples from official artwork and animation.
Recommended CFG: 6.6-6.5
Recommended weights: 0.8-1
Recommended prompt: choker, hat, short hair, twintails,
Honestly, I would have liked to add more detail to the character, but I had a hard time finding a more detailed costume design in the animation. And, because Escha has a tail-like decoration, it was difficult to generate while training and drawing. I will try to solve this problem in a later version by removing some of the tags.
NOTE: The first version is good for generating avatar/portrait/upper body, possibly, there may be a problem in producing full body images
chareschatwintails
Characters-
-Nino
-Miku
-Ichika
-Yotsuba
-Itsuki
Styles-
-Season1
-season2
1000~ Screenshots from the Anime
18,000 Steps/50 Epoch
Trained on NAI
Weights: I didnt really experiment, go ahead
Characters-
-Nino
-Miku
-Ichika
-Yotsuba
-Itsuki
Styles-
-Season1
-season2
1000~ Screenshots from the Anime
18,000 Steps/50 Epoch
Trained on NAI
Weights: I didnt really experiment, go ahead
Nino, 1 girl, solo, pink hair, butterfly hair ornamentIchika, 1 girl, solo, pink hair, short hair, big breastsMiku, 1 girl, solo, brown hair, long hairYotsuba, 1 girl, solo, short hair, yellow hair, green hair ribbonItsuki, 1 girl, solo, red hair, star hair ornament, long hairSeason1Season2
A Lora of jasmine from pokemon.
Use Lora weight of 0.7
Doesn't always get her clothes right but it does get her general appearance right.
A Lora of jasmine from pokemon.
Use Lora weight of 0.7
Doesn't always get her clothes right but it does get her general appearance right.
Jasmine /(pokemon/)
Marie Kanker, from Ed, Edd n Eddy. Everything tagged Marie Kanker and solo on Rule34. Base model was easter_e17.
Marie Kanker, from Ed, Edd n Eddy. Everything tagged Marie Kanker and solo on Rule34. Base model was easter_e17.
Marie KankerEd Edd n Eddy
¿Looking for Momo nsfw version?
**Important read**
This model is a early version, Gradually I will add more pictures and updating the model until make it better and more complete
Hello here some settings that can help to get some good results
Recommended resolution: 512x768
lora weight: 0.70-0.85
Simpling method: DPM++ SDE Karras or Euler a
Steps: 30
Model: "Chilloutmix_NiPrunedFp32Fix" or "Chilloutmix"
CFG: 7
You can use some trigger words like: twmomo, momoring
also you can use: "blonde hair", "brown hair" to get some variations
¿Looking for Momo nsfw version?
**Important read**
This model is a early version, Gradually I will add more pictures and updating the model until make it better and more complete
Hello here some settings that can help to get some good results
Recommended resolution: 512x768
lora weight: 0.70-0.85
Simpling method: DPM++ SDE Karras or Euler a
Steps: 30
Model: "Chilloutmix_NiPrunedFp32Fix" or "Chilloutmix"
CFG: 7
You can use some trigger words like: twmomo, momoring
also you can use: "blonde hair", "brown hair" to get some variations
twmomomomoring
lixiaolang, head lxl, 
From Senki Zesshou Symphogear.
From Senki Zesshou Symphogear.
Kohinata_Miku
Don't be too harsh on me, I'm in the process of learning how to create Lora 🙁
This model was trained on 65 images, with 5 repetitions, a batch size of 2, and 10 epochs, equivalent to 1600 steps.
The model in question that was used was: animefull-final-pruned-fp16.safetensors
I don't recommend a specific weight, try out whichever you like best. I used between 0.6-0.7 for my examples.
-brandish
-fairy tail
-anime coloring
collar, chain, cape, nail polish, fur trim, coat, choker, black choker, cross, hair ornament, cross earings, navel, cleavage. (look at the sample images).
Intalallation:
Option 1: To install it, download it and add it to your "Lora" folder in Stable Diffusion (stable-diffusion-webui\models\Lora).
Option 2: If you are in Collab you can add a code box place this command:
#Add LORA in the Cloud
!wget https://civitai.com/api/download/models/44279 -O /content/stable-diffusion-webui/models/Lora/brandishFairyTail_v10.safetensors
#Add LORA in the Drive
!wget https://civitai.com/api/download/models/44279 -O /content/drive/MyDrive/sd/stable-diffusion-webui/models/Lora/brandishFairyTail_v10.safetensors
Model:
Regarding recommended models, I really like Anything-v4.5 with the VAE anything-v4.0.vae.pt
Finally, you can leave me any suggestions in the comments, they are greatly appreciated 😊💜💚
Don't be too harsh on me, I'm in the process of learning how to create Lora 🙁
This model was trained on 65 images, with 5 repetitions, a batch size of 2, and 10 epochs, equivalent to 1600 steps.
The model in question that was used was: animefull-final-pruned-fp16.safetensors
I don't recommend a specific weight, try out whichever you like best. I used between 0.6-0.7 for my examples.
-brandish
-fairy tail
-anime coloring
collar, chain, cape, nail polish, fur trim, coat, choker, black choker, cross, hair ornament, cross earings, navel, cleavage. (look at the sample images).
Intalallation:
Option 1: To install it, download it and add it to your "Lora" folder in Stable Diffusion (stable-diffusion-webui\models\Lora).
Option 2: If you are in Collab you can add a code box place this command:
#Add LORA in the Cloud
!wget https://civitai.com/api/download/models/44279 -O /content/stable-diffusion-webui/models/Lora/brandishFairyTail_v10.safetensors
#Add LORA in the Drive
!wget https://civitai.com/api/download/models/44279 -O /content/drive/MyDrive/sd/stable-diffusion-webui/models/Lora/brandishFairyTail_v10.safetensors
Model:
Regarding recommended models, I really like Anything-v4.5 with the VAE anything-v4.0.vae.pt
Finally, you can leave me any suggestions in the comments, they are greatly appreciated 😊💜💚
brandish fairy tailanime coloring
Trained on NAI.
Recommended weigth: 1
Current version is not very flexible but it should be enough for now.
feel free to share your result on the comments.
Trained on NAI.
Recommended weigth: 1
Current version is not very flexible but it should be enough for now.
feel free to share your result on the comments.
RoxanneDesir
youxiu
enna alouette from nijisanji en, train with 100 images
enna alouette from nijisanji en, train with 100 images
enna
chinese sexy zombie
chinese sexy zombie
jssn, a woman wearing a hat and a scarfjssn, a woman in a costume with a hat and a swordjssn, a woman in a costume is posing for a picturejssn, a woman with a hat and a costume onjssn, a woman in a blue dress and a hatjssn, a woman in a costume is posing for a picture 
Diviyansha Kaushik is an Indian actress who primarily works in Telugu films. She's one of the prettiest women I've ever seen. This LoRA is trained on around 350 images of her.
Best resolution 768x768 and 768x1024 with a strength of 0.9, however, other resolutions should work fine as well.
Diviyansha Kaushik is an Indian actress who primarily works in Telugu films. She's one of the prettiest women I've ever seen. This LoRA is trained on around 350 images of her.
Best resolution 768x768 and 768x1024 with a strength of 0.9, however, other resolutions should work fine as well.
divi_k
Shamelessly stolen from MalcolmRey(https://civitai.com/user/malcolmrey).
I saw he hadn't converted this one, and since I prefer LORAs so far, decided to just do it. Made even easier that he included his training data in his original checkpoint post.
I took his training images, added a few including nudity, and away we go.
Looking at 0.9-1.1 for strengths
Shamelessly stolen from MalcolmRey(https://civitai.com/user/malcolmrey).
I saw he hadn't converted this one, and since I prefer LORAs so far, decided to just do it. Made even easier that he included his training data in his original checkpoint post.
I took his training images, added a few including nudity, and away we go.
Looking at 0.9-1.1 for strengths
rosariodawson
This is a Lora of the internet celebrity Belle Delphine.
There already exist multiple LoRAs and checkpoints of her, like here, here or here.
But I wasn’t happy with their results, as they either do not look how I would want it to look, or they are very prone to cat ears or pink hair. But everyone has their own taste, so check them out if you are unhappy with this one.
This LoRA does not use any trigger word, and I suggest a default weight of 1.
The preview images mainly use ChilloutMix, additional images are from Deliberate and Realistic Vision. The Overwatch images are using NeverEnding Dream.
The results are good enough for me to upload it here, but it is not perfect, and you might encounter the following problems:
It really likes generating the same lip shape, despite it not even being that frequent in the training images.
This also generates cat ears, especially if you have something with ear in your prompt or something on the head which could become ears, but it is less frequent than with other models in my experiments and opinion.
Depending on the prompt certain noise / “paint” on the face might appear.
Additionally, this LoRA is mostly doing its job only when the face is sufficiently big (also dependent on model). If it becomes too small, the quality decreases and I advise using the high resolution fix then. In the example images multiple images also make use of that.
This Lora is trained on the 1.5 base model with ~140 images, 50 repeats per image, 1e-5 learning rate, 5e-05 for the text encoder, 0.00015 for the unet. All training images were cropped to 512x512 and no bucketing was used.
This is a Lora of the internet celebrity Belle Delphine.
There already exist multiple LoRAs and checkpoints of her, like here, here or here.
But I wasn’t happy with their results, as they either do not look how I would want it to look, or they are very prone to cat ears or pink hair. But everyone has their own taste, so check them out if you are unhappy with this one.
This LoRA does not use any trigger word, and I suggest a default weight of 1.
The preview images mainly use ChilloutMix, additional images are from Deliberate and Realistic Vision. The Overwatch images are using NeverEnding Dream.
The results are good enough for me to upload it here, but it is not perfect, and you might encounter the following problems:
It really likes generating the same lip shape, despite it not even being that frequent in the training images.
This also generates cat ears, especially if you have something with ear in your prompt or something on the head which could become ears, but it is less frequent than with other models in my experiments and opinion.
Depending on the prompt certain noise / “paint” on the face might appear.
Additionally, this LoRA is mostly doing its job only when the face is sufficiently big (also dependent on model). If it becomes too small, the quality decreases and I advise using the high resolution fix then. In the example images multiple images also make use of that.
This Lora is trained on the 1.5 base model with ~140 images, 50 repeats per image, 1e-5 learning rate, 5e-05 for the text encoder, 0.00015 for the unet. All training images were cropped to 512x512 and no bucketing was used.
Makes characters in the style of the Disgaea games.
How to use
Just prompt away. The dataset is based on official character art with "simple background" and "white background" so use those tags if you want to strengthen the style further. You can also add standard tag for Disgaea to further change the style provided the model has data on it, see sample image for how this works on AOM2.
Don't forget to review and upload your own images, it's always fun to see what you're creating!
Like my LoRAs and want more? Please consider a small tip towards my electricity bill and further training >.<
Misc
512 resolution trained on NAI. Sample images from AOM2_NSFW and NAI.
All images are untouched raw txt2imgs, no inpainting done, for quality transparency.
Makes characters in the style of the Disgaea games.
How to use
Just prompt away. The dataset is based on official character art with "simple background" and "white background" so use those tags if you want to strengthen the style further. You can also add standard tag for Disgaea to further change the style provided the model has data on it, see sample image for how this works on AOM2.
Don't forget to review and upload your own images, it's always fun to see what you're creating!
Like my LoRAs and want more? Please consider a small tip towards my electricity bill and further training >.<
Misc
512 resolution trained on NAI. Sample images from AOM2_NSFW and NAI.
All images are untouched raw txt2imgs, no inpainting done, for quality transparency.
jsut test for ai generte
jsut test for ai generte
ai comicschool girljk
v0.1 is suitable for gay621
Please use Hires. fix ,R-ESRGAN Upscaler (Latent is also a good choice)
The LoRA Block Weight extension can be used to make this model only effective in terms of character appearance or coloring
V0.2 is suitable for anything v5, aom3, and counterfeit v2.5, so it may also be suitable for other models
v0.1 is suitable for gay621
Please use Hires. fix ,R-ESRGAN Upscaler (Latent is also a good choice)
The LoRA Block Weight extension can be used to make this model only effective in terms of character appearance or coloring
V0.2 is suitable for anything v5, aom3, and counterfeit v2.5, so it may also be suitable for other models
Weight: ~0.6
Tested dimensions:
W600 * H900
Weight: ~0.6
Tested dimensions:
W600 * H900
ninoblonde_hair, blue_eyes, track_jacket
DO NOT ENTER private area
Ruin exploration is harmful for your body
and today, RUINS are MONITORED by sensors
It was traind with photographs of actual Japanese ruins. These were taken 20 years ago, now those ruins had been vanished.
This model will be improved for your support. See Donations:
Recommended LoRA weight is 0.4 to 0.5.
Watch out for your downloded file name. Write same name into prompts.
<lora:factoryRuins_v10:0.5>, (factory ruins:1.1)Location
indoors, outdoors
Scenenary
train station : impressive broken roofs
railroad tracks, ground vehicle
industrial pipe : chemical plants
indoors, window : inside a room
machine : equipments
steel : rusted structures
building, stone wall
More Apocalyptic
rubble, grass
See also sample image's prompts.
Experimental. Effects depend on your checkpoint.
<lora:factoryRuins_v10:0.5>, (night factory ruins:1.1),(night:1.4)Negative Prompts: dayAnything V3 or later
Realistic models also available, but the quality is less than anime models.
Training based model is ACertainty.
Not for corporate activities (Give me money)
Prohibit redistribution
Thank you for 6 cup of coffee!(Mar.21,2023)
Coffee to me: https://ko-fi.com/kanaejun
BitCoin: 3H33ahQnpL84S4kSKWxCPzKMoEjtdgMBRy
LiteCoin: MWYA5K8jNo9AoCadvPpoMHs9bCEB23QYuk
kanaejun, Japan
DO NOT ENTER private area
Ruin exploration is harmful for your body
and today, RUINS are MONITORED by sensors
It was traind with photographs of actual Japanese ruins. These were taken 20 years ago, now those ruins had been vanished.
This model will be improved for your support. See Donations:
Recommended LoRA weight is 0.4 to 0.5.
Watch out for your downloded file name. Write same name into prompts.
<lora:factoryRuins_v10:0.5>, (factory ruins:1.1)Location
indoors, outdoors
Scenenary
train station : impressive broken roofs
railroad tracks, ground vehicle
industrial pipe : chemical plants
indoors, window : inside a room
machine : equipments
steel : rusted structures
building, stone wall
More Apocalyptic
rubble, grass
See also sample image's prompts.
Experimental. Effects depend on your checkpoint.
<lora:factoryRuins_v10:0.5>, (night factory ruins:1.1),(night:1.4)Negative Prompts: dayAnything V3 or later
Realistic models also available, but the quality is less than anime models.
Training based model is ACertainty.
Not for corporate activities (Give me money)
Prohibit redistribution
Thank you for 6 cup of coffee!(Mar.21,2023)
Coffee to me: https://ko-fi.com/kanaejun
BitCoin: 3H33ahQnpL84S4kSKWxCPzKMoEjtdgMBRy
LiteCoin: MWYA5K8jNo9AoCadvPpoMHs9bCEB23QYuk
kanaejun, Japan
factory ruins
Happy easter! (late, I know!)
Copy pasting from our Patreon:
"Welcome to the wearable LORA series. And welcome to the Experimental part of LORAs by my side.
This embedding is a spinoff of our patreon's character Seizo, the mage desert.
It focus on certain ethnics and headgear, in this case turbant. Classic and iconic item.
I have focused so we have diversity when prompting and even going higher values, it remains like that.
How did I train this?
This LORA is trained in 3 steps. Normal training, re-training (there is a free tutorial in our patreon and lora mixing for final product).
More info in the future since this is a thing I'm working on but these are the basics (always changeable as need).
How this works?
These type of LORA are designed to get good results with minimal prompting. You can see them as "lazy"
LORAs but, they also work with larger prompt, however if you want to get "lazy" you will need to do few things. Of course this lora works on any custom model (1.5 based).
You can either use Clip Skip 1 or 2. Clip Skip 2 will make AI to get more freedom in the generation.
Settings for inferencing:
For painterly style (lazy) (very good for make upscaling and inpainting)
CFG Scale = bellow 5, from 2.5-5 is good value.
Steps = Above 40 necessary (on Euler Ancestral you can try less)
Sampler = Euler Ancestral for variability and Karras Samplers for technical details
For normal style derivative to the model (normal)
CFG Scale = 7 and you can go above.
Steps = 20-30 are good values, 40 could work too.
Sampler = Same as above but SDE works good here.
How to prompt?
This is a "lazy" LORA. It works the same way of other LORAs you may have seen but this lora due to a LORA custom merge can extend it's possibilities:
- It can get good results with just 1 or 2 words. Since it's a wearable with turbant stylish you just have to put man/male woman/female and words like turbant (you can also choose color).
If you use the painterly technique remember to drop way way more than normal CFG Scale and up the steps. You can also use Hi-res Fix scaling to 1, denoise 0.15-0.20 to maintain images a bit cleaner.
- You will get variability even at higher LORA values (over 1 and till 1.5 from my tests) but you can also go 0.X values.
Auxiliar Lora that I recommend even it's not necessary: Epinikion's Offset Noise
This LORA is good for...
Portraits, standing situations, black background images, dark images, high contrast,
normal actions like drinking, sitting, etc. and MIXING with other character LORAs
This LORA is NOT good for
Any sort of landscape. Is a wearable, is not designed for this.
Putting your character in any special situation. It suffers when you don't want to do portraiture or medium standing situations but if you tweak values from 1.2 to 0.8 probably it will be fixed."
If you liked this content please consider support us in our patreon (https://www.patreon.com/user?u=90661954) for more quality material so we can continue doing tutorials, workflow and embeddings.
PS: All of our models, embeddings, etc. are free-use. We don't trust in license and don't want to limit open sourcing.
Happy easter! (late, I know!)
Copy pasting from our Patreon:
"Welcome to the wearable LORA series. And welcome to the Experimental part of LORAs by my side.
This embedding is a spinoff of our patreon's character Seizo, the mage desert.
It focus on certain ethnics and headgear, in this case turbant. Classic and iconic item.
I have focused so we have diversity when prompting and even going higher values, it remains like that.
How did I train this?
This LORA is trained in 3 steps. Normal training, re-training (there is a free tutorial in our patreon and lora mixing for final product).
More info in the future since this is a thing I'm working on but these are the basics (always changeable as need).
How this works?
These type of LORA are designed to get good results with minimal prompting. You can see them as "lazy"
LORAs but, they also work with larger prompt, however if you want to get "lazy" you will need to do few things. Of course this lora works on any custom model (1.5 based).
You can either use Clip Skip 1 or 2. Clip Skip 2 will make AI to get more freedom in the generation.
Settings for inferencing:
For painterly style (lazy) (very good for make upscaling and inpainting)
CFG Scale = bellow 5, from 2.5-5 is good value.
Steps = Above 40 necessary (on Euler Ancestral you can try less)
Sampler = Euler Ancestral for variability and Karras Samplers for technical details
For normal style derivative to the model (normal)
CFG Scale = 7 and you can go above.
Steps = 20-30 are good values, 40 could work too.
Sampler = Same as above but SDE works good here.
How to prompt?
This is a "lazy" LORA. It works the same way of other LORAs you may have seen but this lora due to a LORA custom merge can extend it's possibilities:
- It can get good results with just 1 or 2 words. Since it's a wearable with turbant stylish you just have to put man/male woman/female and words like turbant (you can also choose color).
If you use the painterly technique remember to drop way way more than normal CFG Scale and up the steps. You can also use Hi-res Fix scaling to 1, denoise 0.15-0.20 to maintain images a bit cleaner.
- You will get variability even at higher LORA values (over 1 and till 1.5 from my tests) but you can also go 0.X values.
Auxiliar Lora that I recommend even it's not necessary: Epinikion's Offset Noise
This LORA is good for...
Portraits, standing situations, black background images, dark images, high contrast,
normal actions like drinking, sitting, etc. and MIXING with other character LORAs
This LORA is NOT good for
Any sort of landscape. Is a wearable, is not designed for this.
Putting your character in any special situation. It suffers when you don't want to do portraiture or medium standing situations but if you tweak values from 1.2 to 0.8 probably it will be fixed."
If you liked this content please consider support us in our patreon (https://www.patreon.com/user?u=90661954) for more quality material so we can continue doing tutorials, workflow and embeddings.
PS: All of our models, embeddings, etc. are free-use. We don't trust in license and don't want to limit open sourcing.
man, male, woman, female, turbant
v2.0 2023.04.13
I think it's easier to get a distinctive haircut or pink mesh in the hair.
I suggest you include hair-related keywords.
Recommend options
LoRA weight 0.6~0.8
Trigger words
base
Cure-Spicy
Keywords for more detailed costumes:
tied hair,multicolored hair,side ponytail,streaked hair,blue headwear,pantyhose,blue legwea,blue footwear,
puffy sleeves,white gloves,navy blue footwear,brooch,jewelry,bangs,frills,flower,blue skirt,long skirt,high heels,jewelry,
Try different combinations of words and coordination!
Use animefull based model
Settings
DPM++ Series(SDE Karras, 2M Karras, etc.)
about 20 steps, CFG scale 3.5~6.5
kl-f8-anime2.vae
CLIP skip = 2
Use highres. fix to get higher quality image
Upscaler : Latent series
Denoising strength 0.50~0.65
LoRA training info
trained on sd-scripts by kohya_ss. Thank you a lot!
109 images of Cure-Spicy
replaced character feature tags with Cure-Spicy
resolution : 512x512
rank=alpha=16
batch size = 4, 7 epochs trained
This LoRA may be compatible with animefull-based models (ex. AbyssOrangeMix Series...)
This LoRA can generate nsfw images of Cure-Spicy
This LoRA can not generate nsfw images of huwakokokne
The above text is using a translator
キュアスパイシー（デリシャスパーティプリキュア）
設定などは上記を参考
芙羽ここねは出力できません
v2.0 2023.04.13
I think it's easier to get a distinctive haircut or pink mesh in the hair.
I suggest you include hair-related keywords.
Recommend options
LoRA weight 0.6~0.8
Trigger words
base
Cure-Spicy
Keywords for more detailed costumes:
tied hair,multicolored hair,side ponytail,streaked hair,blue headwear,pantyhose,blue legwea,blue footwear,
puffy sleeves,white gloves,navy blue footwear,brooch,jewelry,bangs,frills,flower,blue skirt,long skirt,high heels,jewelry,
Try different combinations of words and coordination!
Use animefull based model
Settings
DPM++ Series(SDE Karras, 2M Karras, etc.)
about 20 steps, CFG scale 3.5~6.5
kl-f8-anime2.vae
CLIP skip = 2
Use highres. fix to get higher quality image
Upscaler : Latent series
Denoising strength 0.50~0.65
LoRA training info
trained on sd-scripts by kohya_ss. Thank you a lot!
109 images of Cure-Spicy
replaced character feature tags with Cure-Spicy
resolution : 512x512
rank=alpha=16
batch size = 4, 7 epochs trained
This LoRA may be compatible with animefull-based models (ex. AbyssOrangeMix Series...)
This LoRA can generate nsfw images of Cure-Spicy
This LoRA can not generate nsfw images of huwakokokne
The above text is using a translator
キュアスパイシー（デリシャスパーティプリキュア）
設定などは上記を参考
芙羽ここねは出力できません
Cure-Spicy
This is a Dadaptation test. Please make a better lora for her. I have tried a few things and i'm very unsatisfied with the results, lol.
Tags: (bubble_girl:1.2),1girl,blue skin,short hair,blue hair,colored skin,blue skin,diving mask,
Additional tags: diving mask,crop top,underboob,black gloves,midriff,shorts
-Problems:
-diving mask is very inconsistent;
-AOM has problems with colored skins;
-Using the additional tags will cause problems with some position-specific concepts. Make sense, right? Specifying 'shorts' on a lora that has inherently the upper body showing will confuse the model.
This is a Dadaptation test. Please make a better lora for her. I have tried a few things and i'm very unsatisfied with the results, lol.
Tags: (bubble_girl:1.2),1girl,blue skin,short hair,blue hair,colored skin,blue skin,diving mask,
Additional tags: diving mask,crop top,underboob,black gloves,midriff,shorts
-Problems:
-diving mask is very inconsistent;
-AOM has problems with colored skins;
-Using the additional tags will cause problems with some position-specific concepts. Make sense, right? Specifying 'shorts' on a lora that has inherently the upper body showing will confuse the model.
bubble_girl
We will comply with the 774inc. secondary creation guidelines.
We will comply with the 774inc. secondary creation guidelines.
inaba_haneru
A Lora in the style of Nezulet.
Use a Lora weight of 1
A Lora in the style of Nezulet.
Use a Lora weight of 1
Barded Steed
Give it some try, not a carefully trained Lora; you know, not many people are interested in armor and weapon stuff, and you were most likely to find things about 'horse p***s' when you searched the tag 'horse' on this site.
It literally taught AI a concept that horse can also wear armor, you can trigger it with 'barded_steed', and horseback_riding(if you want only a horse, then put it in negative and add no_humans to postive); recommended weight between 0.6 and 0.9, depending on your model
披甲战马
自用Lora，不一定好用；用于给马披上盔甲， 触发关键词：barded_steed, horseback_riding(单画马的话就放negative里,在positive里加no_humans), horse；权重取决于你的模型，一般0.6-0.9
Barded Steed
Give it some try, not a carefully trained Lora; you know, not many people are interested in armor and weapon stuff, and you were most likely to find things about 'horse p***s' when you searched the tag 'horse' on this site.
It literally taught AI a concept that horse can also wear armor, you can trigger it with 'barded_steed', and horseback_riding(if you want only a horse, then put it in negative and add no_humans to postive); recommended weight between 0.6 and 0.9, depending on your model
披甲战马
自用Lora，不一定好用；用于给马披上盔甲， 触发关键词：barded_steed, horseback_riding(单画马的话就放negative里,在positive里加no_humans), horse；权重取决于你的模型，一般0.6-0.9
barded_steed
Note: If you like it, please give it a STAR MARKS !
,
She is a new heroine introduced in Puyo Puyo Chronicle.
Many of you may have been captivated by her love.
The details of the modeling are subtle, but I think the figure has been generally well reproduced, including the coloring.
,
If it doesn't work, the only way to deal with it is to just do the numbers.
And don't stick to the sample prompts.
Thank you for your patience, it is much appreciated.
,
,
※ 気に入っていただけたならぜひ★をつけてください！
,
ぷよぷよクロニクルで登場した新ヒロインです。
彼女の愛に包まれて虜になってしまった方も多いでしょう。
細かい造形は微妙ですが、色合い含めておおむね再現できていると思います。
,
うまくいかない場合の対処方法はひたすら数をこなすこと。
そしてサンプルプロンプトにこだわらないことです。
あなたの忍耐に感謝いたします。
,
,
,
<lora:ally:0.8> kawaii ally cute 1girl, green eyes, purple pendant, light brown hair, medium hair, yellow pantyhose, green dress, lightgreen shirt, green boots naughty smile style parody
,
Negative prompt
badhandv4 bad_prompt EasyNegative, 2girls NSFW
,
Steps: 20, Sampler: DPM++ SDE Karras, CFG scale: 4, Seed: 3117124254, Size: 384x512, Model hash: 55b3fdcfc6, Model: caulkinumV2_AR, Denoising strength: 0.6, Clip skip: 2, ENSD: 31337, Hires upscale: 2, Hires upscaler: R-ESRGAN 4x+ Anime6B
,
,
,
,
caulkinumV2_AR
https://civitai.com/models/32537/caulkinumv2ar-for-anime-style
You can use it to highquality illustrations.
クオリティの高いイラストを錬成できるので使ってみてください。
,
,
Coffee will increase work efficiency and motivation.
コーヒーは作業効率を高めモチベーションをアップさせるでしょう。
,
Thank you for reading to the end.
I hope one of you will use it.
,
最後まで読んでくれてありがとうございます。
使用してくれるか方が一人でもいることを祈っています。
Note: If you like it, please give it a STAR MARKS !
,
She is a new heroine introduced in Puyo Puyo Chronicle.
Many of you may have been captivated by her love.
The details of the modeling are subtle, but I think the figure has been generally well reproduced, including the coloring.
,
If it doesn't work, the only way to deal with it is to just do the numbers.
And don't stick to the sample prompts.
Thank you for your patience, it is much appreciated.
,
,
※ 気に入っていただけたならぜひ★をつけてください！
,
ぷよぷよクロニクルで登場した新ヒロインです。
彼女の愛に包まれて虜になってしまった方も多いでしょう。
細かい造形は微妙ですが、色合い含めておおむね再現できていると思います。
,
うまくいかない場合の対処方法はひたすら数をこなすこと。
そしてサンプルプロンプトにこだわらないことです。
あなたの忍耐に感謝いたします。
,
,
,
<lora:ally:0.8> kawaii ally cute 1girl, green eyes, purple pendant, light brown hair, medium hair, yellow pantyhose, green dress, lightgreen shirt, green boots naughty smile style parody
,
Negative prompt
badhandv4 bad_prompt EasyNegative, 2girls NSFW
,
Steps: 20, Sampler: DPM++ SDE Karras, CFG scale: 4, Seed: 3117124254, Size: 384x512, Model hash: 55b3fdcfc6, Model: caulkinumV2_AR, Denoising strength: 0.6, Clip skip: 2, ENSD: 31337, Hires upscale: 2, Hires upscaler: R-ESRGAN 4x+ Anime6B
,
,
,
,
caulkinumV2_AR
https://civitai.com/models/32537/caulkinumv2ar-for-anime-style
You can use it to highquality illustrations.
クオリティの高いイラストを錬成できるので使ってみてください。
,
,
Coffee will increase work efficiency and motivation.
コーヒーは作業効率を高めモチベーションをアップさせるでしょう。
,
Thank you for reading to the end.
I hope one of you will use it.
,
最後まで読んでくれてありがとうございます。
使用してくれるか方が一人でもいることを祈っています。
To achieve optimal results, consider incorporating both LoRA v.1.1 and v.1.0.
By blending these two models with the recommended weights of 0.45 for LoRA v.1.1 and 0.55 for LoRA v.1.0.
Look at the example image of LoRA v.1.1, which showcases the exceptional results that can be achieved by merging the two versions with the recommended weight.
These are simply suggestions and you are free to tailor them to your personal preferences.
*
And I apologize for any inconvenience, but I am unsure how to merge LoRA. Is there any way to merge them? Thank you.
••••••••••••••••••••••••••••••••••••••••••••••••••••••
To achieve the unique and captivating style of Hajime Sorayama, try using the trigger words "sorayama style".
••••••••••••••••••••••••••••••••••••••••••••••••••••••
I've made my very first LoRa. I'm still getting the hang of it, so any advice or recommendations you have would be incredibly valuable to me. And, of course, don't forget to share your artwork here! Thanks for your support and let's have some fun!
To achieve optimal results, consider incorporating both LoRA v.1.1 and v.1.0.
By blending these two models with the recommended weights of 0.45 for LoRA v.1.1 and 0.55 for LoRA v.1.0.
Look at the example image of LoRA v.1.1, which showcases the exceptional results that can be achieved by merging the two versions with the recommended weight.
These are simply suggestions and you are free to tailor them to your personal preferences.
*
And I apologize for any inconvenience, but I am unsure how to merge LoRA. Is there any way to merge them? Thank you.
••••••••••••••••••••••••••••••••••••••••••••••••••••••
To achieve the unique and captivating style of Hajime Sorayama, try using the trigger words "sorayama style".
••••••••••••••••••••••••••••••••••••••••••••••••••••••
I've made my very first LoRa. I'm still getting the hang of it, so any advice or recommendations you have would be incredibly valuable to me. And, of course, don't forget to share your artwork here! Thanks for your support and let's have some fun!
Sorayama Style
Does it actually say "blacked"? Well, mostly not really. You know SD can't really handle text. Maybe some inpainting and post edit would do the trick.
This is a simple outfit LoRA. Similarly to my other LoRAs, I try to make it so that the effect of the LoRA has the least impact as possible on the character, style and composition. You can use weight above 1.0. See example below.
It can be used with various types of clothing but the basics are: sports bra and thong.
With my Piper Perri Meme LoRA and now this Blacked underwear LoRA, can you guess what's coming next? Be sure to follow to get updates.
Please rate if you like it - your feedback lets me know you enjoy my work!
Does it actually say "blacked"? Well, mostly not really. You know SD can't really handle text. Maybe some inpainting and post edit would do the trick.
This is a simple outfit LoRA. Similarly to my other LoRAs, I try to make it so that the effect of the LoRA has the least impact as possible on the character, style and composition. You can use weight above 1.0. See example below.
It can be used with various types of clothing but the basics are: sports bra and thong.
With my Piper Perri Meme LoRA and now this Blacked underwear LoRA, can you guess what's coming next? Be sure to follow to get updates.
Please rate if you like it - your feedback lets me know you enjoy my work!
sports brathong
Hey, I recently started reading "The Strongest Florist" and I thought I'd create a couple of models from this series. This is the first one :)
https://www.webtoons.com/en/action/the-strongest-florist
I think it turned out quite well - feel free to use it.
Hey, I recently started reading "The Strongest Florist" and I thought I'd create a couple of models from this series. This is the first one :)
https://www.webtoons.com/en/action/the-strongest-florist
I think it turned out quite well - feel free to use it.
may_the_strongest_florist
recommended weight: 0.7-0.9
trigger words: long pony tail, up pony tail, white cape, black short pants
recommended weight: 0.7-0.9
trigger words: long pony tail, up pony tail, white cape, black short pants
long pony tailup ponu tailwhite capeblack short pants
This model works great for inpainting leg warmers on your character! pairs well with chilloutmix and chilled_re-generic-v2 and v3. I have not tested it with an anime model so try at your own risk.
Follow me for more developments :) : https://twitter.com/AikaKittie
This model works great for inpainting leg warmers on your character! pairs well with chilloutmix and chilled_re-generic-v2 and v3. I have not tested it with an anime model so try at your own risk.
Follow me for more developments :) : https://twitter.com/AikaKittie
leg warmerssocksknee socks
Fai, Rona, Seira, Yuusha, Mao and Mei from the anime series Endro~!
(Mu mu mu mu mu mu~ sold separately)
How to use
Use the character trigger word, usually that's enough but in some cases you might need to add one or two extra tags describing them.
Due to the number of characters and variance in the dataset they're not all evenly balanced, some are slightly more stubborn about changing outfits but if you drop down either the LoRA strength or the character tag strength it should sort itself out.
Don't forget to review and upload your own images, it's always fun to see what you're creating!
Like my LoRAs and want more? Please consider a small tip towards my electricity bill and further training >.<
Misc
512 resolution trained on NAI. Sample images from AOM2_NSFW
All images are untouched raw txt2imgs, no inpainting done, for quality transparency.
The wildcard prompt below can be used to "gatcha" random characters or used partly to prompt specific characters (or block them if one is spilling over to others with your model)
<lora:Endro:0.8>,
(1girl)(solo)
{
Fai \(Endro!\),blonde hair, yellow eyes,hair ornament, hair flowers|
Mao \(Endro!\), red hair, demon wings, curled horns|
Mei \(Endro!\), blue hair,green eyes|
Rona \(Endro!\),blonde hair, hair curls, pink eyes|
Seira \(Endro!\), pointy ears, very long purple hair,hair band, hair flowers|
Yuusha \(Endro!\), pink hair, long hair, bangs, blue eyes, ahoge
}
Fai, Rona, Seira, Yuusha, Mao and Mei from the anime series Endro~!
(Mu mu mu mu mu mu~ sold separately)
How to use
Use the character trigger word, usually that's enough but in some cases you might need to add one or two extra tags describing them.
Due to the number of characters and variance in the dataset they're not all evenly balanced, some are slightly more stubborn about changing outfits but if you drop down either the LoRA strength or the character tag strength it should sort itself out.
Don't forget to review and upload your own images, it's always fun to see what you're creating!
Like my LoRAs and want more? Please consider a small tip towards my electricity bill and further training >.<
Misc
512 resolution trained on NAI. Sample images from AOM2_NSFW
All images are untouched raw txt2imgs, no inpainting done, for quality transparency.
The wildcard prompt below can be used to "gatcha" random characters or used partly to prompt specific characters (or block them if one is spilling over to others with your model)
<lora:Endro:0.8>,
(1girl)(solo)
{
Fai \(Endro!\),blonde hair, yellow eyes,hair ornament, hair flowers|
Mao \(Endro!\), red hair, demon wings, curled horns|
Mei \(Endro!\), blue hair,green eyes|
Rona \(Endro!\),blonde hair, hair curls, pink eyes|
Seira \(Endro!\), pointy ears, very long purple hair,hair band, hair flowers|
Yuusha \(Endro!\), pink hair, long hair, bangs, blue eyes, ahoge
}
Fai \(Endro!\)Mao \(Endro!\)Mei \(Endro!\)Rona \(Endro!\)Seira \(Endro!\)Yuusha \(Endro!\)
Another character from Mega Man series, this time Leviathan from Mega Man Zero! :D
Hope you like it!
Recommended prompting
1girl, solo, blue eyes, android, helmet
Another character from Mega Man series, this time Leviathan from Mega Man Zero! :D
Hope you like it!
Recommended prompting
1girl, solo, blue eyes, android, helmet
leviathan_megamanz
Making LoRA's is a thing I do for fun. So if you like what I do? You can help me out by just using the LoRA and leaving a review!
I like Sphynx cats, so I made a LoRA about Sphynx cats. Okay now for the nitty and gritty. This LoRA allows you to create photorealistic Sphynx cats. It also depends of course on the checkpoint you use. But e.g. in Realistic Vision the Sphynx cats look really good, but from my testing it can work on some anime based checkpoints too (some anime checkpoints can make cursed images though 💀), it's just better at realistic checkpoints. Be sure to check out my uploaded images, I've made sure to use a variety of checkpoints.
The LoRA weights that seem to give the best results are between 0.7 and 0.9, 1.0 also works, but I've noticed that it's then sometimes harder to e.g. prompt eye or skin colour. The trigger word for this LoRA is sphynx cat, these additional prompts might help you out:
legs and paws might help you getting their legs and paws in the image.
yellow eyes, blue eyes, green eyes should work, since used dataset has Sphynx cat with different eye colours.
tail might help you getting their tail in the image, it might also help you force a perspective where their tail is in the image.
full body might help you getting a Sphynx cat fully in the image.
looking at the camera and looking to the side might give a certain angle the Sphynx cat will look at.
eating out of a bowl of food might help you get an image where the Sphynx cat eats out of food bowl.
looking up might help you get the Sphynx cat to look up.
I've tagged the skin colours of the Sphynx cats in the dataset, so you might have luck using the following prompts: white skin, gray and white skin, white and spotted black skin, gray skin, black skin, brown skin, black and white skin, brown and white skin, spotted black and brown with white skin & white and brown skin.
wearing a collar might help you to get the Sphynx cat to wear a collar.
wearing a sweater and some variants on it should help you to get the Sphynx cat to wear a sweater, wearing a shirt might also work.
standing, laying & sitting might help you get the Sphynx cat in those positions.
Making LoRA's is a thing I do for fun. So if you like what I do? You can help me out by just using the LoRA and leaving a review!
I like Sphynx cats, so I made a LoRA about Sphynx cats. Okay now for the nitty and gritty. This LoRA allows you to create photorealistic Sphynx cats. It also depends of course on the checkpoint you use. But e.g. in Realistic Vision the Sphynx cats look really good, but from my testing it can work on some anime based checkpoints too (some anime checkpoints can make cursed images though 💀), it's just better at realistic checkpoints. Be sure to check out my uploaded images, I've made sure to use a variety of checkpoints.
The LoRA weights that seem to give the best results are between 0.7 and 0.9, 1.0 also works, but I've noticed that it's then sometimes harder to e.g. prompt eye or skin colour. The trigger word for this LoRA is sphynx cat, these additional prompts might help you out:
legs and paws might help you getting their legs and paws in the image.
yellow eyes, blue eyes, green eyes should work, since used dataset has Sphynx cat with different eye colours.
tail might help you getting their tail in the image, it might also help you force a perspective where their tail is in the image.
full body might help you getting a Sphynx cat fully in the image.
looking at the camera and looking to the side might give a certain angle the Sphynx cat will look at.
eating out of a bowl of food might help you get an image where the Sphynx cat eats out of food bowl.
looking up might help you get the Sphynx cat to look up.
I've tagged the skin colours of the Sphynx cats in the dataset, so you might have luck using the following prompts: white skin, gray and white skin, white and spotted black skin, gray skin, black skin, brown skin, black and white skin, brown and white skin, spotted black and brown with white skin & white and brown skin.
wearing a collar might help you to get the Sphynx cat to wear a collar.
wearing a sweater and some variants on it should help you to get the Sphynx cat to wear a sweater, wearing a shirt might also work.
standing, laying & sitting might help you get the Sphynx cat in those positions.
sphynx cat
This is my first ever LoRA. Hope you enjoy it!
LoRA model for the character Wattson from Apex Legends. Trained for 20 epochs on the Hassaku checkpoint model.
LoRA weight can be anywhere from 0.1-1. It should work fine at 1.
CFG scale 6-8, although you can try experimenting.
Trigger word (optional): "Wattson"
This LoRA is very NSFW friendly, but it can do SFW images just fine as well.
Sometimes multiple subjects/fractals may appear but it is rare for this to happen, and will get ironed out over the next couple updates.
How to use LoRA's in automatic1111's web-ui:
Update webui (use git pullor redownload it)
Copy the file to stable-diffusion-webui/models/lora
Select your LoRA like in this video
Make sure to select your desired weight, or leave it at the default of 1. (example: <lora:WattsonLora:1> is 1 weight, <lora:WattsonLora:0.5> is 0.5 weight)
If you like this LoRA, I would appreciate a heart as thanks :)
This is my first ever LoRA. Hope you enjoy it!
LoRA model for the character Wattson from Apex Legends. Trained for 20 epochs on the Hassaku checkpoint model.
LoRA weight can be anywhere from 0.1-1. It should work fine at 1.
CFG scale 6-8, although you can try experimenting.
Trigger word (optional): "Wattson"
This LoRA is very NSFW friendly, but it can do SFW images just fine as well.
Sometimes multiple subjects/fractals may appear but it is rare for this to happen, and will get ironed out over the next couple updates.
How to use LoRA's in automatic1111's web-ui:
Update webui (use git pullor redownload it)
Copy the file to stable-diffusion-webui/models/lora
Select your LoRA like in this video
Make sure to select your desired weight, or leave it at the default of 1. (example: <lora:WattsonLora:1> is 1 weight, <lora:WattsonLora:0.5> is 0.5 weight)
If you like this LoRA, I would appreciate a heart as thanks :)
wattson
有使用问题可以在 b站视频 内留言。
带 prompt 图片可以在公众号【老梵AI】后台回复【雅妃模型】获取
本人所有图片和模型均免费公开，如果有人向你收费，可以举报他。
主模型：chilloutmix_Ni
依赖 Lora：koreanDollLikeness_v15
其他参数可参考图片 demo
建议权重：
雅妃：0.5-0.6
koreanDoll：0.5-0.6
有使用问题可以在 b站视频 内留言。
带 prompt 图片可以在公众号【老梵AI】后台回复【雅妃模型】获取
本人所有图片和模型均免费公开，如果有人向你收费，可以举报他。
主模型：chilloutmix_Ni
依赖 Lora：koreanDollLikeness_v15
其他参数可参考图片 demo
建议权重：
雅妃：0.5-0.6
koreanDoll：0.5-0.6
无
有使用问题可以在 b站视频 内留言。
带 prompt 图片可以在公众号【老梵AI】后台回复【女武神模型】获取
本人所有图片和模型均免费公开，如果有人向你收费，可以举报他。
主模型：chilloutmix_Ni
依赖 Lora：koreanDollLikeness_v15
其他参数可参考图片 demo
建议权重：
云曦：0.5-0.6
koreanDoll：0.5-0.6
有使用问题可以在 b站视频 内留言。
带 prompt 图片可以在公众号【老梵AI】后台回复【女武神模型】获取
本人所有图片和模型均免费公开，如果有人向你收费，可以举报他。
主模型：chilloutmix_Ni
依赖 Lora：koreanDollLikeness_v15
其他参数可参考图片 demo
建议权重：
云曦：0.5-0.6
koreanDoll：0.5-0.6
无
Trained with Anything-v3 model.
Recommended weight is 0.8, but play around a bit!
Supports NSFW generations.
Trained with Anything-v3 model.
Recommended weight is 0.8, but play around a bit!
Supports NSFW generations.
mizuhara
A request from a Kofi donator
Lewis from Girls' Frontline
Keyword: Lewis
Don't ask her what happened in 1932
Recommend 0.8-1 strength
If you've enjoyed my models, consider leaving a tip.
https://ko-fi.com/leontalon
A request from a Kofi donator
Lewis from Girls' Frontline
Keyword: Lewis
Don't ask her what happened in 1932
Recommend 0.8-1 strength
If you've enjoyed my models, consider leaving a tip.
https://ko-fi.com/leontalon
Lewis
🐱 Her purry voice rings a familiar tune.
Works at weight 1. Made with my Lora guide.
🐱 Her purry voice rings a familiar tune.
Works at weight 1. Made with my Lora guide.
mikenekoblack dress, bandaged arm, white thighhighsheterochromia, purple eyes, red eyeseyepatchstuffed animal
This is my first uploaded LoRA. I really only made it to see what would happen since the images from the show are so stylized, I thought it would never work.
It was trained using Kohya_ss with v1.5 and regularization images. Only the girls are included and sometimes need a little prompt help if you want to stay true to each character, otherwise it will mix hair and eye colors etc.
This is my first uploaded LoRA. I really only made it to see what would happen since the images from the show are so stylized, I thought it would never work.
It was trained using Kohya_ss with v1.5 and regularization images. Only the girls are included and sometimes need a little prompt help if you want to stay true to each character, otherwise it will mix hair and eye colors etc.
BlossomBubblesButtercup
This is a LORA of the character Anarchy Stocking from the anime PSG.
To prompt this character in her angel outfit, enter "<lora:Stocking-PSG:strength>Anarchy Stocking, off shoulder corset dress, white striped thighhighs, hair bow, white choker, sandals,". To prompt this character in her gothic outfit, enter "<lora:Stocking-PSG:strength>Anarchy Stocking, gothic frilled dress, black striped thighhighs, hair bow, mary janes, ". Accessories include "striped sword, mini wings, angel wings, halo, stuffed toy, lollipop,". Prompting "lightshow background," will prompt the transformation sequence background from the show. There's also a few other small outfits trained in here as well.
This LORA looks the best at 0.6 ~ 0.8 strength in my opinion. Stocking was trained on Anythingv4.5 at Clip Skip 2. This lora works great across other models too.
This is a LORA of the character Anarchy Stocking from the anime PSG.
To prompt this character in her angel outfit, enter "<lora:Stocking-PSG:strength>Anarchy Stocking, off shoulder corset dress, white striped thighhighs, hair bow, white choker, sandals,". To prompt this character in her gothic outfit, enter "<lora:Stocking-PSG:strength>Anarchy Stocking, gothic frilled dress, black striped thighhighs, hair bow, mary janes, ". Accessories include "striped sword, mini wings, angel wings, halo, stuffed toy, lollipop,". Prompting "lightshow background," will prompt the transformation sequence background from the show. There's also a few other small outfits trained in here as well.
This LORA looks the best at 0.6 ~ 0.8 strength in my opinion. Stocking was trained on Anythingv4.5 at Clip Skip 2. This lora works great across other models too.
off shoulder corset dresswhite striped thighhighswhite chokersandalsgothic dressblack striped thighhighsmary janeshair bowhaloangel wingsmini wingsstuffed toystriped swordlightshow background
The model this LoRA is extracted from, A to Zovya RPG Artist's Tools (1.5 & 2.1) was created to encompass all things RPG Fantasy Art generation related. From landscapes, ruins, castles to characters and monster concepts. It's designed to assist artist wanting to incorporate AI into their art workflow.
If you'd like to have a smaller file instead of the Checkpoint, you can use this LoRA alongside the base 1.5 model or another of your choice to get a similar style.
Do you have requests? I've been putting in many more hours lately with this. That's my problem, not yours. But if you'd like to tip me, buy me a beer. Beer encourages me to ignore work and make AI models instead. Tip and make a request. I'll give it a shot if I can. Here at Ko-Fi
The model this LoRA is extracted from, A to Zovya RPG Artist's Tools (1.5 & 2.1) was created to encompass all things RPG Fantasy Art generation related. From landscapes, ruins, castles to characters and monster concepts. It's designed to assist artist wanting to incorporate AI into their art workflow.
If you'd like to have a smaller file instead of the Checkpoint, you can use this LoRA alongside the base 1.5 model or another of your choice to get a similar style.
Do you have requests? I've been putting in many more hours lately with this. That's my problem, not yours. But if you'd like to tip me, buy me a beer. Beer encourages me to ignore work and make AI models instead. Tip and make a request. I'll give it a shot if I can. Here at Ko-Fi
zrpgstyle
loRA of Ryo from "Bochi the Rock" anime
loRA of Ryo from "Bochi the Rock" anime
Ryo
loRA that copies Heijun's style
loRA that copies Heijun's style
This is a LoRa with the style of the Akebi-chan no Sailor-fuku anime.
The base model is Suzumehachi.
This is a LoRa with the style of the Akebi-chan no Sailor-fuku anime.
The base model is Suzumehachi.
akebi
A LoRA for creating photorealistic Futa TakerPOV blowjob images.
If you feel like supporting my work, please consider visiting my Patreon to get access to more detailed instructions on how to get the most out of my models: Link.
Update: Version 2 was trained with adjusted training settings and is less overfittet. Image quality should be much better now.
This one is pretty straight forward to use.
Use 'takerpovbj' as the general trigger.
'Close up', 'extreme close up', 'looking up' and 'looking ahead' are used to control the perspective, but you don't need to use any of them.
Also reacts well to 'small/medium/big/huge breasts'.
It was also trained to react to 'grabbing you', but I didn't have much luck prompting for it.
Something to work out with the next version.
That's about it. Looking forward to seeing what you can make with it!
Recommended strength is around 0.7-0.8.
Based on UberRealisticPornMerge
If you like this you might also like my other LoRAs.
A LoRA for creating photorealistic Futa TakerPOV blowjob images.
If you feel like supporting my work, please consider visiting my Patreon to get access to more detailed instructions on how to get the most out of my models: Link.
Update: Version 2 was trained with adjusted training settings and is less overfittet. Image quality should be much better now.
This one is pretty straight forward to use.
Use 'takerpovbj' as the general trigger.
'Close up', 'extreme close up', 'looking up' and 'looking ahead' are used to control the perspective, but you don't need to use any of them.
Also reacts well to 'small/medium/big/huge breasts'.
It was also trained to react to 'grabbing you', but I didn't have much luck prompting for it.
Something to work out with the next version.
That's about it. Looking forward to seeing what you can make with it!
Recommended strength is around 0.7-0.8.
Based on UberRealisticPornMerge
If you like this you might also like my other LoRAs.
takerpovbjlooking uplooking aheadclose upextreme close upgrabbing you
Lighter loRA of Emma from Nikke Godness of Victory game
Lighter loRA of Emma from Nikke Godness of Victory game
Emma
We will comply with the 774inc. secondary creation guidelines.
We will comply with the 774inc. secondary creation guidelines.
suou_patra
Only Hammann appears! And before the Retrofit ! ! !
This is LoRA. LoHa and DB are uploaded at almost the same time, so please be careful not to make a mistake.
Some adjustments are required, but all fit well.
AOM
7th layer
HeavenOrangeMix
kawaiDiffusionSD15_v4Nsight
linneamix
paleyMIX
Animelike Series (no longer exists in Civitai) [DELETED]
sita7taker
hammann_\(azur_lane\),1girl
The words are the same as LoRA and LoHa, so I will use the same explanation!
Basically, the higher the percentage, the stronger the influence.
"solo" and "white_background" are also 100% triggers, but you should use them depending on the situation. Because "white_background" is strange even though the background is the ocean, isn't it?
Q. What is version AdaFactor...?
The scheduler is different when learning.
Lion did not learn well 😭
Q. Is it necessary to divide into DB, LoRA, and LoHa?
Since the number of learning is very small, it was necessary to make up for it by branching 😭
Also, LoRA and LoHa are not perfect. DB+LoRA is a good option for those who don't know "LoRA Block Weight". Of course, for advanced users, LoHa may be enough!
Q. Ret...
No Retrofit Hammann! ! It's not in the dataset!
Q. Where is the Retrofit version?
Not now 😭 I might make one someday, I might not make it 😑
Only Hammann appears! And before the Retrofit ! ! !
This is LoRA. LoHa and DB are uploaded at almost the same time, so please be careful not to make a mistake.
Some adjustments are required, but all fit well.
AOM
7th layer
HeavenOrangeMix
kawaiDiffusionSD15_v4Nsight
linneamix
paleyMIX
Animelike Series (no longer exists in Civitai) [DELETED]
sita7taker
hammann_\(azur_lane\),1girl
The words are the same as LoRA and LoHa, so I will use the same explanation!
Basically, the higher the percentage, the stronger the influence.
"solo" and "white_background" are also 100% triggers, but you should use them depending on the situation. Because "white_background" is strange even though the background is the ocean, isn't it?
Q. What is version AdaFactor...?
The scheduler is different when learning.
Lion did not learn well 😭
Q. Is it necessary to divide into DB, LoRA, and LoHa?
Since the number of learning is very small, it was necessary to make up for it by branching 😭
Also, LoRA and LoHa are not perfect. DB+LoRA is a good option for those who don't know "LoRA Block Weight". Of course, for advanced users, LoHa may be enough!
Q. Ret...
No Retrofit Hammann! ! It's not in the dataset!
Q. Where is the Retrofit version?
Not now 😭 I might make one someday, I might not make it 😑
hammann_\(azur_lane\)1girl
Providing coffees or commissions start here!
The awaited Ruby Model is here! I feel like I have to put this out here, but this model does tend to produce NSFW work. You've been warned :)
Providing coffees or commissions start here!
The awaited Ruby Model is here! I feel like I have to put this out here, but this model does tend to produce NSFW work. You've been warned :)
xyzruby
Li Wen (WILL: A Wonderful World) 李雯 (WILL：美好世界)
Weight has to be 0.6-0.8.
This LoRA is an experiment to answer whether I can make a decent LoRA with only copies of the same image. I will probably do this experiment with all the characters from WILL: A Wonderful World.
For Alicia, the dataset was this:
Used Cetus-mix and Kotosmix for example images.
General tags:
Young Alicia: youngalicia_will, 1girl, long hair, dark skin, dark-skinned female, wavy hair, brown hair, purple eyes
Adult Alicia: adultalicia_will, 1girl, dark skin, dark-skinned female, brown hair, long hair, large breasts, multicolored hair, gradient hair
Outfits:
Young Alicia: young_defaultoutfit, necklace, bracelet, bare shoulders, purple skirt, hoop earrings, cross, purple hairband, collarbone, off-shoulder shirt, purple shirt, anklet, purple footwear
Adult Alicia: adult_defaultoutfit, fingerless gloves, cleavage, jewelry, earrings, elbow gloves, pantyhose, collarbone, holster, black gloves, choker, thigh strap, high heels, black footwear, high heel boots
Don't bother using it at a weight higher than 0.8, it'll overfit.
Some models can't quite replicate her adult clothes.
"ss_text_encoder_lr":"1e-05",
"ss_total_batch_size":"2",
"ss_num_epochs":"10",
"ss_unet_lr":"0.0001",
"ss_learning_rate":"0.0001",
"ss_lr_scheduler":"cosine_with_restarts",
"ss_lr_warmup_steps":"0",
"ss_max_token_length":"225",
"ss_mixed_precision":"fp16",
"ss_network_alpha":"64.0",
"ss_network_dim":"128",
{\"n_repeats\": 20, \"img_count\": 10}
"ss_resolution":"(768, 768)"
"ss_sd_model_name":"Animefull-final-pruned.ckpt",
"ss_seed":"42",
Li Wen (WILL: A Wonderful World) 李雯 (WILL：美好世界)
Weight has to be 0.6-0.8.
This LoRA is an experiment to answer whether I can make a decent LoRA with only copies of the same image. I will probably do this experiment with all the characters from WILL: A Wonderful World.
For Alicia, the dataset was this:
Used Cetus-mix and Kotosmix for example images.
General tags:
Young Alicia: youngalicia_will, 1girl, long hair, dark skin, dark-skinned female, wavy hair, brown hair, purple eyes
Adult Alicia: adultalicia_will, 1girl, dark skin, dark-skinned female, brown hair, long hair, large breasts, multicolored hair, gradient hair
Outfits:
Young Alicia: young_defaultoutfit, necklace, bracelet, bare shoulders, purple skirt, hoop earrings, cross, purple hairband, collarbone, off-shoulder shirt, purple shirt, anklet, purple footwear
Adult Alicia: adult_defaultoutfit, fingerless gloves, cleavage, jewelry, earrings, elbow gloves, pantyhose, collarbone, holster, black gloves, choker, thigh strap, high heels, black footwear, high heel boots
Don't bother using it at a weight higher than 0.8, it'll overfit.
Some models can't quite replicate her adult clothes.
"ss_text_encoder_lr":"1e-05",
"ss_total_batch_size":"2",
"ss_num_epochs":"10",
"ss_unet_lr":"0.0001",
"ss_learning_rate":"0.0001",
"ss_lr_scheduler":"cosine_with_restarts",
"ss_lr_warmup_steps":"0",
"ss_max_token_length":"225",
"ss_mixed_precision":"fp16",
"ss_network_alpha":"64.0",
"ss_network_dim":"128",
{\"n_repeats\": 20, \"img_count\": 10}
"ss_resolution":"(768, 768)"
"ss_sd_model_name":"Animefull-final-pruned.ckpt",
"ss_seed":"42",
youngalicia_willadultalicia_will
Sea Angels in the style of Sashizume Soutarou. If you don't know what this means consider hitting the back button and download something else, this LoRA can put out some serious nightmare fuel true to its source material.
How to use
Trigger word "sengle", suggested additional tags - "colored skin", "grey skin", "third eye", "tentacle hair", "sharp teeth", "monster girl"
Don't forget to review and upload your own images, it's always fun traumatizing to see what you're creating!
Like my LoRAs and want more? Please consider a small tip towards my electricity bill and further training >.<
Misc
512 resolution trained on NAI. Sample images from AOM2_NSFW
All images are untouched raw txt2imgs, no inpainting done, for quality transparency.
Sea Angels in the style of Sashizume Soutarou. If you don't know what this means consider hitting the back button and download something else, this LoRA can put out some serious nightmare fuel true to its source material.
How to use
Trigger word "sengle", suggested additional tags - "colored skin", "grey skin", "third eye", "tentacle hair", "sharp teeth", "monster girl"
Don't forget to review and upload your own images, it's always fun traumatizing to see what you're creating!
Like my LoRAs and want more? Please consider a small tip towards my electricity bill and further training >.<
Misc
512 resolution trained on NAI. Sample images from AOM2_NSFW
All images are untouched raw txt2imgs, no inpainting done, for quality transparency.
sengle
First LoRA
First LoRA
sajo_yukimiidolmaster idolmaster_cinderella_girls 
Reiko Date Wingfield from the Fault!! eroge visual novel
2 versions; 8 epochs and 9 epochs
ALL, MIDD, OUTD, and OUTALL LoRA block weights give good results.
Use with 0.7+ weights
Reiko Date Wingfield from the Fault!! eroge visual novel
2 versions; 8 epochs and 9 epochs
ALL, MIDD, OUTD, and OUTALL LoRA block weights give good results.
Use with 0.7+ weights
reiko_fault
a simple lora based on xretakex's art style https://twitter.com/xretakex
anything v4.5 checkpoint, anything v4.0 vae, clip skip 2, weight 1, euler_a
let me know what you think!
a simple lora based on xretakex's art style https://twitter.com/xretakex
anything v4.5 checkpoint, anything v4.0 vae, clip skip 2, weight 1, euler_a
let me know what you think!
Requested by and sources provided by @ericiscool.
I changed my training parameters for this one so it might perform slightly differently than before. However, I manage to reduce the size to 72MB for file size.
Hires steps:17, denoising:0.5+
She might look like someone from real life, but she is not what you think. She is a totally fictional character.
It is prohibited to use this model for commercial purposes and any scenarios of illegal acts and purposes.
本人原则上不支持使用此模型生成任何成人内容。用户在使用本模型时，请务必遵守当地法律法规，勿侵犯他人的名誉权、隐私权或肖像权。
Requested by and sources provided by @ericiscool.
I changed my training parameters for this one so it might perform slightly differently than before. However, I manage to reduce the size to 72MB for file size.
Hires steps:17, denoising:0.5+
She might look like someone from real life, but she is not what you think. She is a totally fictional character.
It is prohibited to use this model for commercial purposes and any scenarios of illegal acts and purposes.
本人原则上不支持使用此模型生成任何成人内容。用户在使用本模型时，请务必遵守当地法律法规，勿侵犯他人的名誉权、隐私权或肖像权。
Shinohara Mari from Taimanin PRGX
train by using Anything-v3
"aged up" for Bandit (future version)
Shinohara Mari from Taimanin PRGX
train by using Anything-v3
"aged up" for Bandit (future version)
Shinohara Mari
neteromeruem
Making models can be expensive. Do you like what I do? Consider supporting me on Patreon 🅿️ or feel free to buy me a coffee ☕
Another difficult one as she has a very odd outfit and the dataset wasn't really too much consistent. In the end I found the right combination to get her accurately enough.
Suggested weight is 0.5 or 0.6 and triggers with kaine \(nier\) and her outfit/hair description. This lora wil ladd all those concepts to your anime models. Examples have been made with Anything v4.5 and Counterfeit V2.
How to use LoRA's in auto1111:
Update webui (use git pull like here or redownload it)
Copy the file to stable-diffusion-webui/models/lora
Select your LoRA like in this video
Make sure to change the weight (by default it's :1 which is usually too high)
Making models can be expensive. Do you like what I do? Consider supporting me on Patreon 🅿️ or feel free to buy me a coffee ☕
Another difficult one as she has a very odd outfit and the dataset wasn't really too much consistent. In the end I found the right combination to get her accurately enough.
Suggested weight is 0.5 or 0.6 and triggers with kaine \(nier\) and her outfit/hair description. This lora wil ladd all those concepts to your anime models. Examples have been made with Anything v4.5 and Counterfeit V2.
How to use LoRA's in auto1111:
Update webui (use git pull like here or redownload it)
Copy the file to stable-diffusion-webui/models/lora
Select your LoRA like in this video
Make sure to change the weight (by default it's :1 which is usually too high)
kaine \(nier\)
Consider supporting me on Patreon for more models and content.
Anime style of Rent-a-Girlfriend (Kanojo, Okarishimasu). You can produce the characters from the anime (only females) with the right prompts.
(This is a requested LoRA in one of my comments section that I decided to fulfill)
This isn't the best for NSFW, because the anime doesn't show anything, but you can force it anyway.
Try tags that has to do with their appeareance. If you want an specific character you should type their names like below.
Examples:
chizuru_ichinose, bangs, long hair, brown hair
mami_nanami, short hair, bangs, green_eyes, blonde hair, hair_between_eyes, multiclolored_hair
ruka_sarashina, short_hair, bangs, bow, green_eyes, hair_bow, (blue_bow)
List of names for you: chizuru_ichinose, mami_nanami, ruka_sarashina, sumi_sakurasawa
Used orangemix VAE for most images
I'm not using the official lora loader, so you should use the additional-network webui plugin to use my lora setting.
Trained using ambientmixAnAnime_v10, but should be ok with similar models. Weight should be near 0.9-1.0 (used 1.0 most of the time)
Please credit me when using any of my models.
Specially when merging this model.
You cannot use my models in other generation websites.
Consider supporting me on Patreon for more models and content.
Anime style of Rent-a-Girlfriend (Kanojo, Okarishimasu). You can produce the characters from the anime (only females) with the right prompts.
(This is a requested LoRA in one of my comments section that I decided to fulfill)
This isn't the best for NSFW, because the anime doesn't show anything, but you can force it anyway.
Try tags that has to do with their appeareance. If you want an specific character you should type their names like below.
Examples:
chizuru_ichinose, bangs, long hair, brown hair
mami_nanami, short hair, bangs, green_eyes, blonde hair, hair_between_eyes, multiclolored_hair
ruka_sarashina, short_hair, bangs, bow, green_eyes, hair_bow, (blue_bow)
List of names for you: chizuru_ichinose, mami_nanami, ruka_sarashina, sumi_sakurasawa
Used orangemix VAE for most images
I'm not using the official lora loader, so you should use the additional-network webui plugin to use my lora setting.
Trained using ambientmixAnAnime_v10, but should be ok with similar models. Weight should be near 0.9-1.0 (used 1.0 most of the time)
Please credit me when using any of my models.
Specially when merging this model.
You cannot use my models in other generation websites.
chizuru_ichinosemami_nanamiruka_sarashinasumi_sakurasawa
Use Negative Prompts! Try lower LoRA Strengh
All Preview Images are 0.4 Strenght!
Generate Big LoRA Pack cost many Time and Training ... so pleace write a 5* review when you support me for free
angela
elena
haruka
katia
mari
maya
satsuki
shizuno
sophia
tokiko
vasilisa
"name" - see list
anime art style - Trigger for Anime Style (Beta)
See Preview Images for my Prompts.
LoRA Load over: additional-Networks (work without it)
Model: abyssorangemix2_Hard
VAE: Anything-V3.0.vae.pt
Txt2Img and Upscale settings:
DPM++ SDE Karras
Steps: 20
480x720
CFG: 7
LoRA Strengh: 0.4
Upscale: 2x
Denoise: 0.6
Upscaler: "R-ESRGAN 4x+ Anime6B"
Use Negative Prompts! Try lower LoRA Strengh
All Preview Images are 0.4 Strenght!
Generate Big LoRA Pack cost many Time and Training ... so pleace write a 5* review when you support me for free
angela
elena
haruka
katia
mari
maya
satsuki
shizuno
sophia
tokiko
vasilisa
"name" - see list
anime art style - Trigger for Anime Style (Beta)
See Preview Images for my Prompts.
LoRA Load over: additional-Networks (work without it)
Model: abyssorangemix2_Hard
VAE: Anything-V3.0.vae.pt
Txt2Img and Upscale settings:
DPM++ SDE Karras
Steps: 20
480x720
CFG: 7
LoRA Strengh: 0.4
Upscale: 2x
Denoise: 0.6
Upscaler: "R-ESRGAN 4x+ Anime6B"
angelaelenaharukakatiamarimayasatsukishizunosophiatokikovasilisa
A Character LoRA for everyone's favorite genocidal robot popstar from the ratchet and clank games! Works best at a strength of 0.85 and a prompt that includes at least "courtneygears, robotgirl, brown hair, orange eyes, hairband, yellow sclera". Somewhat imposes it's artstyle on the image so you may want to tweak any additional artstyle LORA strengths. Check the example images for prompts that work well. Can do softcore NSFW but the training material didn't include any robot genitals. Trained on 38 good images, of which a few are direct blender renders of the model T-posing. Trained on nai, examples generated with AnythingV4.5
This is the second time i'm making a LoRA and this time i've learned how to do it mostly properly, but there's still room for improvement. If i have enough free time i'd love to release some improved versions.
A Character LoRA for everyone's favorite genocidal robot popstar from the ratchet and clank games! Works best at a strength of 0.85 and a prompt that includes at least "courtneygears, robotgirl, brown hair, orange eyes, hairband, yellow sclera". Somewhat imposes it's artstyle on the image so you may want to tweak any additional artstyle LORA strengths. Check the example images for prompts that work well. Can do softcore NSFW but the training material didn't include any robot genitals. Trained on 38 good images, of which a few are direct blender renders of the model T-posing. Trained on nai, examples generated with AnythingV4.5
This is the second time i'm making a LoRA and this time i've learned how to do it mostly properly, but there's still room for improvement. If i have enough free time i'd love to release some improved versions.
courtneygears
Making models can be expensive. Do you like what I do? Consider supporting me on Patreon 🅿️ or feel free to buy me a coffee ☕
Trained on AnyLoRA, works great on it and on most AOM/Anything derived anime models. Use au \(d elete\) for more accuracy and weight should be around 1. For the upscaler I like the effect it gets using Latent at 0.6 denoise.
How to use LoRA's in auto1111:
Update webui (use git pull like here or redownload it)
Copy the file to stable-diffusion-webui/models/lora
Select your LoRA like in this video
Make sure to change the weight according to the instructions (by default it's :1)
Making models can be expensive. Do you like what I do? Consider supporting me on Patreon 🅿️ or feel free to buy me a coffee ☕
Trained on AnyLoRA, works great on it and on most AOM/Anything derived anime models. Use au \(d elete\) for more accuracy and weight should be around 1. For the upscaler I like the effect it gets using Latent at 0.6 denoise.
How to use LoRA's in auto1111:
Update webui (use git pull like here or redownload it)
Copy the file to stable-diffusion-webui/models/lora
Select your LoRA like in this video
Make sure to change the weight according to the instructions (by default it's :1)
au \(d elete\)
"Sendou Aichi" is the main character of the "Cardfight!! Vanguard"
"Sendou Aichi" is the main character of the "Cardfight!! Vanguard"
Sendou_Aichi
LoRA trained on Lea.
Trigger word is leaCode.
Her eyes are somewhat pink-ish so I recommend using "red eyes". Also, her hair is alterable so use "ponytail" to get the standard hairstyle.
To get her full outfit use "pauldrons", "detached red sleeves", "bracers", "grey shorts".
Her scar is sometimes barely noticeable, but it's usually there.
Weight around ~1 should be fine.
Trained on NAI
LoRA trained on Lea.
Trigger word is leaCode.
Her eyes are somewhat pink-ish so I recommend using "red eyes". Also, her hair is alterable so use "ponytail" to get the standard hairstyle.
To get her full outfit use "pauldrons", "detached red sleeves", "bracers", "grey shorts".
Her scar is sometimes barely noticeable, but it's usually there.
Weight around ~1 should be fine.
Trained on NAI
leaCode
Penis on face, penis on mouth, cheek press, cheek poking, penis between eyes, 2girls, 3girls, 4girls, 5girls
Trained on 'penis on face, penis on X,' as the kept tags for everything.
penis on face: main tag
penis on mouth: penis covering mouth
cheek press: penis pressing into cheeck
cheek poking: penis poking into cheek
facejob: penis rubbing against face; works particularly well with 2girls, 3girls, etc...
penis between eyes: penis in center of face, this is actually what inspired me to train this
2girls: yep
3girls: tipple yep
4girls: quadruple yep!
5girls: quintuple yep!
6+girls: no. apparently yes? it's just a bit gacha
This one works well, but it's kinda hard to control some aspects. I tried something new with this, and the results are promising, but mixed.
Penis on face, penis on mouth, cheek press, cheek poking, penis between eyes, 2girls, 3girls, 4girls, 5girls
Trained on 'penis on face, penis on X,' as the kept tags for everything.
penis on face: main tag
penis on mouth: penis covering mouth
cheek press: penis pressing into cheeck
cheek poking: penis poking into cheek
facejob: penis rubbing against face; works particularly well with 2girls, 3girls, etc...
penis between eyes: penis in center of face, this is actually what inspired me to train this
2girls: yep
3girls: tipple yep
4girls: quadruple yep!
5girls: quintuple yep!
6+girls: no. apparently yes? it's just a bit gacha
This one works well, but it's kinda hard to control some aspects. I tried something new with this, and the results are promising, but mixed.
penis on facepenis on mouthcheek presscheek pokingpenis between eyes2girls3girls4girls5girls
Reposting it since the website didn't want to upload my model due to bug of browser...
Whatever weight pleases you, trigge'r is Kirari.
Trained with novelai best result with it but you can use others models.
Give me your feedback~
Reposting it since the website didn't want to upload my model due to bug of browser...
Whatever weight pleases you, trigge'r is Kirari.
Trained with novelai best result with it but you can use others models.
Give me your feedback~
Kirari
This is a style model that gives softer faces to characters, this was a mix of several artist that although not doing what I originally intended turned out very cute. Tends to work best at 0.6. This model was trained on NAI anime-full-pruned and generated in AOM3. Made with the Kohya tool for AI Training through google colab. Any suggestions to improve or recommended things for the data set would be appreciated.
This is a style model that gives softer faces to characters, this was a mix of several artist that although not doing what I originally intended turned out very cute. Tends to work best at 0.6. This model was trained on NAI anime-full-pruned and generated in AOM3. Made with the Kohya tool for AI Training through google colab. Any suggestions to improve or recommended things for the data set would be appreciated.
Tips to use:
Use LoRA weight between 0.6-0.8
To summon her use the following tags: white heart, 1girl, blue hair, red eyes, symbol-shaped pupils, short hair with long locks, power symbol, ahoge, bare shoulders, leotard, elbow gloves, white bodysuit
If you have any questions feel free to ask!
Hope you guys like it! ❤️
You can ask for character LoRAs on our patreon page. Consider support us. 😘
Tips to use:
Use LoRA weight between 0.6-0.8
To summon her use the following tags: white heart, 1girl, blue hair, red eyes, symbol-shaped pupils, short hair with long locks, power symbol, ahoge, bare shoulders, leotard, elbow gloves, white bodysuit
If you have any questions feel free to ask!
Hope you guys like it! ❤️
You can ask for character LoRAs on our patreon page. Consider support us. 😘
white heart1girlblue hairred eyesshort hair with long locksbare shoulderssymbol-shaped pupilswhite bodysuit
There is a dark wave, electro, gothic nightclub lost in time, long before the advent of social media, when images used to accumulate in galleries and pages of the web 1.0. This checkpoint was trained on images from the age of the first mass-market digital cameras 30 years ago until 15 years later. The source material has been a blurry mess of low-resolution images with lots of JPEG artefacts; they’ve all been pre-processed over several iterations with different AI upscaling and artefact removal as well as face restoration systems.
The training data from 13025 images have been blended during the training process as additions to the already present similar subjects of the SD 1.5 checkpoint. A subset of the images was already in the original low-resolution versions included in SD 1.5. The result is that this checkpoint is trained to visualise epic and surreal dark wave, electro, and gothic party scenes populated with guests and visitors wearing elaborate costumes and unique clothing with the right make-up to go along for the occasion.
The training has been set up to get the general art and a static look and feel of the dataset transferred into a versatile SD 1.5 checkpoint where everything else is preserved to allow for a wide range of subjects and compositions. The checkpoint is not trained to replicate the actual persons or a precise subset of the training data. The training took 50 epochs over 1630240 iterations.
Use “GothClub2K” in your prompts to invoke attention to the training data and dataset of the subjects alongside a detailed description of what you like to visualise.
for CP see https://civitai.com/models/36430/gothic-nightclub-dark-wave-electro-party
There is a dark wave, electro, gothic nightclub lost in time, long before the advent of social media, when images used to accumulate in galleries and pages of the web 1.0. This checkpoint was trained on images from the age of the first mass-market digital cameras 30 years ago until 15 years later. The source material has been a blurry mess of low-resolution images with lots of JPEG artefacts; they’ve all been pre-processed over several iterations with different AI upscaling and artefact removal as well as face restoration systems.
The training data from 13025 images have been blended during the training process as additions to the already present similar subjects of the SD 1.5 checkpoint. A subset of the images was already in the original low-resolution versions included in SD 1.5. The result is that this checkpoint is trained to visualise epic and surreal dark wave, electro, and gothic party scenes populated with guests and visitors wearing elaborate costumes and unique clothing with the right make-up to go along for the occasion.
The training has been set up to get the general art and a static look and feel of the dataset transferred into a versatile SD 1.5 checkpoint where everything else is preserved to allow for a wide range of subjects and compositions. The checkpoint is not trained to replicate the actual persons or a precise subset of the training data. The training took 50 epochs over 1630240 iterations.
Use “GothClub2K” in your prompts to invoke attention to the training data and dataset of the subjects alongside a detailed description of what you like to visualise.
for CP see https://civitai.com/models/36430/gothic-nightclub-dark-wave-electro-party
GothClub2K
Support me training more model . Patreon.com/goofyai
Lora weight 0.7-1 works well.
This model has very less flexibility. Pruned for her default outfit so changing her outfit might be hard so don't bother. I will try to train more flexible model in future if I had time.
drop your review also check my Patreon
check the preview images for prompting guide
Support me training more model . Patreon.com/goofyai
Lora weight 0.7-1 works well.
This model has very less flexibility. Pruned for her default outfit so changing her outfit might be hard so don't bother. I will try to train more flexible model in future if I had time.
drop your review also check my Patreon
check the preview images for prompting guide
kaho(blue archive)
Please use this LoRA and please share your works for everyone to appreciate, thank you
ขอบคุณทุกคน ที่แบ่งปัน เพราะผมได้ความรู้และศึกษาจากที่นี้มาเยอะเลยถือโอกาส แบ่งปันบ้าง
ขอให้ทุกคนสนุกกับการสร้างงานของตัวเอง
ขอบคุณครับ
Please use this LoRA and please share your works for everyone to appreciate, thank you
ขอบคุณทุกคน ที่แบ่งปัน เพราะผมได้ความรู้และศึกษาจากที่นี้มาเยอะเลยถือโอกาส แบ่งปันบ้าง
ขอให้ทุกคนสนุกกับการสร้างงานของตัวเอง
ขอบคุณครับ
Pegasus Seiya
LoRA trained on 768x512 images of Merc G63.
Have tried on SD1.5, Protogen, revAnimated, Dreamshaper, Anything and it works well.
Will need Hires fix for best results, strength - 0.5-0.8
LoRA trained on 768x512 images of Merc G63.
Have tried on SD1.5, Protogen, revAnimated, Dreamshaper, Anything and it works well.
Will need Hires fix for best results, strength - 0.5-0.8
gwag63
Based on @shiindraws art style. Very nsfw.
Based on @shiindraws art style. Very nsfw.
wide hips, huge ass, nude, 1 girl, huge breasts
Now that things have settled backdown on Civitai, I am reopening the LoRA shop of my ai creations. Next up is Sin: Nanatsu no Taizai Mammon, which alongside Luficer has been the most dm request to bring back to my page.
Now that things have settled backdown on Civitai, I am reopening the LoRA shop of my ai creations. Next up is Sin: Nanatsu no Taizai Mammon, which alongside Luficer has been the most dm request to bring back to my page.
mammon_sevenmortalsins
weight:0.8-1
tag:multiple girls, 2girls, yae miko, raiden shogun,
I wrote realistic+anime in the title, because some buddies seem to not know that many Loras are universal...
Yes,you can use nsfw,group sex，That's exactly what I put them together for!
In addition to standard clothing, some special clothing and action tags, please go to the sample picture to find
Of course, you can also call them separately.
我在标题写上了realistic+anime，因为有些哥们好像到现在还不知道很多Lora是通用的……
呵呵，除了双人tag，你也可以用nsfw啦，group sex这些，我就是为了这个才把她俩放一起训练的。
一些特殊的服装和动作，去示例图看吧
当然，你也可以单独叫出她俩
weight:0.8-1
tag:multiple girls, 2girls, yae miko, raiden shogun,
I wrote realistic+anime in the title, because some buddies seem to not know that many Loras are universal...
Yes,you can use nsfw,group sex，That's exactly what I put them together for!
In addition to standard clothing, some special clothing and action tags, please go to the sample picture to find
Of course, you can also call them separately.
我在标题写上了realistic+anime，因为有些哥们好像到现在还不知道很多Lora是通用的……
呵呵，除了双人tag，你也可以用nsfw啦，group sex这些，我就是为了这个才把她俩放一起训练的。
一些特殊的服装和动作，去示例图看吧
当然，你也可以单独叫出她俩
multiple girls, 2girls, yae miko, raiden shogun,
Now that things have settled backdown on Civitai, I am reopening the LoRA shop of my ai creations. Starting with Sin: Nanatsu no Taizai Lucifer. Mature content was used in the making this LoRA, user discretion advised.
Now that things have settled backdown on Civitai, I am reopening the LoRA shop of my ai creations. Starting with Sin: Nanatsu no Taizai Lucifer. Mature content was used in the making this LoRA, user discretion advised.
lucifer_sevenmortalsins
Nabiki Tendo is part of the Ranma 1/2 character collection that you can find on my profile as well as Akane, Shampoo. The Lora is very flexible and works perfectly in various models and styles. Use keywords for greater accuracy. “NabikiWaifu” as well as her school uniform “School Uniform” I personally prefer to use 0.7 strength in the Lora but you can try with the one that works best for you. REMEMBER, if you liked the Lora support me with a review and some image, that motivates me to continue making Loras. Another thing, the character dataset is carefully selected and refined so as not to copy a specific style and thus achieve a more flexible Lora!! :D
Nabiki Tendo is part of the Ranma 1/2 character collection that you can find on my profile as well as Akane, Shampoo. The Lora is very flexible and works perfectly in various models and styles. Use keywords for greater accuracy. “NabikiWaifu” as well as her school uniform “School Uniform” I personally prefer to use 0.7 strength in the Lora but you can try with the one that works best for you. REMEMBER, if you liked the Lora support me with a review and some image, that motivates me to continue making Loras. Another thing, the character dataset is carefully selected and refined so as not to copy a specific style and thus achieve a more flexible Lora!! :D
NabikiWaifuSchool uniformdress
Art style with interesting shading, originally greyscale but does color images too.
How to use
No trigger word needed. Just prompt 1girl or 1boy and go. While trained on solo characters this LoRA turned out fairly versatile, see samples for prompts.
Don't forget to review and upload your own images, it's always fun to see what you're creating!
Like my LoRAs and want more? Please consider a small tip towards my electricity bill and further training >.<
Misc
512 resolution trained on NAI. Sample images from AOM2_NSFW
All images are untouched raw txt2imgs, no inpainting done, for quality transparency.
Art style with interesting shading, originally greyscale but does color images too.
How to use
No trigger word needed. Just prompt 1girl or 1boy and go. While trained on solo characters this LoRA turned out fairly versatile, see samples for prompts.
Don't forget to review and upload your own images, it's always fun to see what you're creating!
Like my LoRAs and want more? Please consider a small tip towards my electricity bill and further training >.<
Misc
512 resolution trained on NAI. Sample images from AOM2_NSFW
All images are untouched raw txt2imgs, no inpainting done, for quality transparency.
A LORA model trained on Fabio Mayor.
To use the model place the trigger words "fabiomayor person" in the prompt.
Trained on 5510 steps.
A LORA model trained on Fabio Mayor.
To use the model place the trigger words "fabiomayor person" in the prompt.
Trained on 5510 steps.
fabiomayor person