These are Lora mix made using my own models. I've used either my models uploaded here on Civitai or models that I trained privately and didn't post due to quality issues. These mix are a way for me to try and make use of some of my less "useful" lora, and hopefully make something worth posting. I will try to give credit to the artists or redirect to the models used as much as possible.-- Soft Anime --This Lora Mix is trained primarily with art from <a target="_blank" rel="ugc" href="https://x.com/fluffylewdsai">@fluffylewdsai</a>, <a target="_blank" rel="ugc" href="https://x.com/Greatm8SFM">@Greatm8SFM</a>, and <a target="_blank" rel="ugc" href="https://x.com/coffeelove68">@coffeelove68</a>This model works best at 1.0Necessary Trigger Word: FluffyLewdsStyle, Greatm8Style,Feel free to include one trigger word and not the other, the style changes but it might be fun to play around with and see what you can get.These tags might also help: CoffeeLoverStyle, greatm8,The CoffeeLover style is mixed pretty weakly into the mix (&gt;25%) and in testing I don't think the tag had much impact on the overall style, but add it if you think it'd help.Pony seems to have a basic understanding of the "greatm8" style, so adding the tag can sometimes help.Update 8/12/24 A lora mix issue within Kohya had applied the wrong hash to the model. This issue should be fixed.-- Clean Anime --I used a prototype version of my <a target="_blank" rel="ugc" href="https://civarchive.com/models/645754/juswa-or-artist-style-or">Juswa </a>style as the base for this mix. The rest of this mix is from my <a target="_blank" rel="ugc" href="https://civarchive.com/models/547289/wet-skin?modelVersionId=608729">Wet Skin</a> lora, as well as a number of other failed lora projects that I didn't feel like they were good enough to post.This model works best at 1.0Necessary Trigger Word: JustWatchStyle, SweatyStyle,Feel free to include one trigger word and not the other, the style changes but it might be fun to play around with and see what you can get.-- Rough Anime --This is a re-post of a model that I had already uploaded a few months ago. However, now that I have a dedicated Lora-Mix page, I felt that it should be uploaded here instead of being tacked on as a V2 to another lora. This Lora heavily uses my <a target="_blank" rel="ugc" href="https://civarchive.com/models/561969">Rough Anime Sketch</a> lora, my <a target="_blank" rel="ugc" href="https://civarchive.com/models/547289?modelVersionId=609775">Wet Skin V2</a> Lora, and my <a target="_blank" rel="ugc" href="https://civarchive.com/models/529886/thick-lines">Thick Lines</a> lora.I usually run it at 1.0, however sometimes poses seem to become real stagnant. If you find all your images posing pretty much the same, lower the strength just a tad to between 0.8-0.9.Necessary Trigger Word: SketchStyleThis lora really likes ADetailer to help fix faces. Also, personally found that the roughness of the colors worked very well with pixelart lora.I am aware of some sort of Hash-related issues that this model has with Citivai and automatic1111. Unfortunately I'm not quite sure how to fix this issue.-- VurDV-Mix --This Lora Mix is trained primarily with art from <a target="_blank" rel="ugc" href="https://x.com/VurDV8a8gbGRctd">@VurDV8a8gbGRctd</a>, <a target="_blank" rel="ugc" href="https://x.com/eggtmt0303">@eggtmt0303</a>, <a target="_blank" rel="ugc" href="https://x.com/PlanV260344">@PlanV260344</a>. These lora were of low quality, and never made it to Civitai, but the mix is pretty good.This model Works best at 1.0Necessary Trigger Word: VurDVStyle, EggtmtStyle, PlanV2Style,These tags might also help: monster girl,The VurDV style (which makes up most of the mix) was heavily trained on monster girl images. This mix does monster girls pretty well, as a general concept, but tends to have a problem understanding different types of monster girls. Other monster girl lora might be useful in this case, but I haven't really tested it.If you like my work and want to give me some extra support, you can always tip⚡Buzz⚡to help me make more lora.

Pengu997  | Lora Mixes | Styles

This Lora Mix is trained primarily with art from <a target="_blank" rel="ugc" href="https://x.com/fluffylewdsai">@fluffylewdsai</a>, <a target="_blank" rel="ugc" href="https://x.com/Greatm8SFM">@Greatm8SFM</a>, and <a target="_blank" rel="ugc" href="https://x.com/coffeelove68">@coffeelove68</a>This model works best at 1.0Necessary Trigger Word: FluffyLewdsStyle, Greatm8Style,Feel free to include one trigger word and not the other, the style changes but it might be fun to play around with and see what you can get.These tags might also help: CoffeeLoverStyle, greatm8,The CoffeeLover style is mixed pretty weakly into the mix (&gt;25%) and in testing I don't think the tag had much impact on the overall style, but add it if you think it'd help.Pony seems to have a basic understanding of the "greatm8" style, so adding the tag can sometimes help.Update 8/12/24 A lora mix issue within Kohya had applied the wrong hash to the model. This issue should be fixed.

Files with SHA256: f8bb0c6299e204c61eb14c7d8e3e62e63408e4a3500f0427e349a7c1ef0d2ce3