/kemono/ - Kemono

Fluffy Anime Girls

Mode: Reply
Name
Subject
Message

Max message length: 20000

Files

Max file size: 30.00 MB

Max files: 5

Captcha
E-mail
Password

(used to delete files and postings)

Misc

Remember to follow the rules

[ Return / Catalog / Bottom ]

(6.88 MB 2800x2048 00150-4052202118.png)
Kemono AI Art Thread Tail Toucher 04/08/2025 (Tue) 12:53:43 No. 2125
(5.49 MB 2048x2800 00143-983682260.png)
This is some starter info from the last thread but with a small edition. >>1632 Getting an AI UI running on your PC has really been streamlined these days. >Requirements: Have a GPU with at least 6GBs of VRAM (not to be confused with normal RAM). If you use Stable Diffusion Forge or ReForge, it's made to run off of weaker hardware so you can still use SDXL models. It's good to aim for 12+ GBs of VRAM. I use a 4070 Super since it's made for AI. It has 12GBs of VRAM and I can gen 1024x1400 images in 13 seconds. AMD cards work but they're more of a pain to setup. https://github.com/LykosAI/StabilityMatrix This setups everything for you in a very simple manner and is constantly updated with better features. It also has access to basically every AI UI you want. I use Reforge. >Extensions There are like 5 extensions I think are mandatory for generation: <Adetailer Automatically inpaints both hands and face while generating. Hand inpainting can by tricky but it works very well on the face. Using this means you can generate an image but inpaint with Adetailer to generate a face with a completely different art style. You can even use another checkpoint just for the Adetailer portion of the gen (this may crash your UI just to warn you). I use this to get the body shape and style from one artist while using another to get a certain kind of eyes. If you just want to have higher quality faces, then just leave the adetailer prompt blank. I recommend setting the steps higher and the pixels from 512x512 to 1024x1024 for a higher quality inpaint. <Dynamic Prompts Allows you to use wildcards. Wildcards are a .txt file with a list of words. When added to a prompt like: Kemono_girls It will randomly select a prompt from that .txt file. Everything in one line is a prompt. Different prompts are separated by a line. So you can have bat girl, bat ears, bat wings, in one line and it will use that whole line as a prompt. Having a dozen of these wildcards can give a lot of diversity to your prompts. It really helps to go on Civit AI to search for them. Here are a few of my old wildcards. https://pixeldrain.com/u/T4w9gD26 <a1111-sd-webui-tagcomplete Another must have extension. It auto completes words based on Danbooru tags. Makes it so much easier to make prompts for characters and tags with this helping you out. There's also an e621 version. Considering that models are based on danbooru tags this will make your prompt much better. <sd-webui-lobe-theme If you're using Forge, Reforge, or A1111 UIs, this upgrades the UI into something much more productive and useful. <sd-webui-regional-prompter Allows for region prompting which means you can set a prompt to show up in a certain area. Gives you way more control on what's being generated. Also makes it much easier to have multiple distinct characters in the same gen without their features being mixed. <Vectorscope / Diffusion Color Grading (Diffusion-CG) Vectorscope controls the brightness and contrast of an image while generating. It's great for making darker images without relying on Vpred models. Diffusion-CG helps with colors while generating. If you think your images are too grey or saturated despite changing your VAE. Try this.
(8.37 MB 2048x2800 00025-4008443464.png)
(7.68 MB 2048x2800 00038-632565901.png)
(7.75 MB 2048x2800 00030-1083587452.png)
Some models I can recommend for generating kemono girls. Unlike SD1.5 models, you can practically use any Noob/Illustrious model to gen kemono girls. Of course those designed strictly for them will do a better job but most of these models can do it just fine. Though you might need to fight with the prompt a little more depending on the model. >SakuraPlatinum https://civitai.com/models/934925/sakuraplatinumillustriousxl?modelVersionId=1523324 >T-NoobNAI3 https://civitai.com/models/823566/t-noobnai3?modelVersionId=1460600 >BoobAI V4.0 https://civitai.com/models/971445/boobai?modelVersionId=1456266 >MAGIC_ILL Vpred V1.0 https://civitai.com/models/1346879/magicill?modelVersionId=1567607 >IcedCoffeeIL V2.0 https://civitai.com/models/1299502?modelVersionId=1598066 >SteinIllustrious https://civitai.com/models/1078146?modelVersionId=1612091 >Nova Anime XL https://civitai.com/models/376130/nova-anime-xl?modelVersionId=1225516 >NTR Mix V4.0/XIII https://civitai.com/models/926443?modelVersionId=1166878 https://civitai.com/models/926443?modelVersionId=1061268 >Mol_Keun Mix Deepblue/V3 https://civitai.com/models/135477?modelVersionId=1524472 https://civitai.com/models/135477?modelVersionId=1579444 >Nova Furry XL V5 https://civitai.com/models/503815?modelVersionId=1513134 There is a Noob version of Yiffymix but I could never get good results for it. Though that could just be me. https://civitai.com/models/3671?modelVersionId=1623867
Copying this from a 4chan post. Basically how to color an image or edit it with features you want. Using reforge or forge or a111 Enlable controlnet Pick the image Pick "lineart" Pick invert from black to white Pick the lineart model (you may need to download it) Make a prompt that suits the image, in this case >anime coloring, 1girl, blue ribbon, blonde hair, green eyes, twintails, freckles, blush, accidental exposure, nipples, cross-laced clothes, purple dress, elbow gloves Generate image. Basically you just generate the image normally but use controlnet to give it the lines Is quite simple
(4.66 MB 1761x2304 00016-3484050815.png)
(3.74 MB 1792x2304 00024-438628850.png)
(4.15 MB 1792x2304 00015-3452434263.png)
(4.18 MB 1792x2304 00018-42991735.png)
(4.24 MB 1792x2304 00021-1948291121.png)
Now with all the pre-recs out the way. Time for some gens. A friend requested Rouge The Bat paizuri. So I made a ton of that.
(3.66 MB 1792x2304 00026-3589685344.png)
(4.26 MB 1792x2304 00032-983329163.png)
(4.08 MB 1792x2304 00031-3662268783.png)
(4.20 MB 1792x2304 00028-191366804.png)
(4.24 MB 1792x2304 00030-3043325808.png)
>>2132 Now in a sweater and shirt.
(4.54 MB 1792x2304 00043-3253370910.png)
(3.63 MB 1792x2304 00038-1672700074.png)
(4.01 MB 1792x2304 00040-1180096498.png)
(4.18 MB 1792x2304 00044-3408147889.png)
(4.34 MB 1792x2304 00049-1239804918.png)
>>2133 Cheerleader outfit.
(3.95 MB 1792x2304 00055-481798365.png)
(4.00 MB 1792x2304 00053-4092971268.png)
(4.27 MB 1792x2304 00056-4244283929.png)
(4.62 MB 1792x2304 00059-3917932484.png)
(4.71 MB 1792x2304 00058-3874184350.png)
>>2134 A more realistic artstyle thanks to Hizake. I'll post more Rouge another time.
(1.33 MB 1288x1008 00046-136873162.png)
(8.12 MB 2048x2600 00035-3123022089.png)
(2.00 MB 1296x1024 00020-4089957190.png)
(2.18 MB 1024x1296 00017-3718542339.png)
Tried to remake an old SD1.5 gen. Came out pretty well this time around.
(5.02 MB 1536x2688 00022-1153293550.png)
(5.90 MB 1536x2688 00018-1640785228.png)
(5.70 MB 1536x2688 00019-267556857.png)
(5.68 MB 1536x2688 00020-207825904.png)
(5.70 MB 1536x2688 00017-3250575735.png)
Attempts at a deer girl and trying out a romper outfit. This style is a combo of Kenichi Sonoda (Gunsmith Cats), and Final Fantasy Tactics.
(6.01 MB 1536x2688 00012-2216244296.png)
(5.77 MB 1536x2688 00027-867866691.png)
(5.44 MB 1536x2688 00026-2193642832.png)
(6.06 MB 2048x2600 00034-3873287912.png)
(6.05 MB 2048x2600 00041-994021808.png)
Took the outfit and applied it to a dragon girl. It's a retake on this gen >>1058 but now taller and more slender. Some of them are more /monster/ than /kemono/.
(6.59 MB 2048x2600 00068-4244717819.png)
(5.66 MB 2048x2600 00077-2514383462.png)
(5.60 MB 2048x2600 00085-824421268.png)
(5.63 MB 2048x2600 00083-1820093901.png)
(6.28 MB 2048x2600 00081-3327819867.png)
>No tall dragon GF
>>2138 That one's a great cross between kemono and monster girl, saying as someone who is into both.
(6.67 MB 2048x2600 00097-922779522.png)
(6.27 MB 2048x2600 00095-837033006.png)
(6.41 MB 2048x2600 00100-249008574.png)
(5.60 MB 2048x2600 00085-824421268.png)
Who wouldn't want to hug a dragon girl? >>2140 Thanks. The main problem I kept running into is that the lora for the artist Hidari would create some horrible background artifacts and ruin some things. Like ruining the tail or background, which is a problem that modern Illustrious mixes shouldn't happen.
(5.10 MB 2048x2048 00126-3322624362.png)
(5.10 MB 2048x2048 00127-2972065373.png)
(5.44 MB 2048x2048 00128-2881343667.png)
A man and his combat doll enjoying some tea together.
(5.38 MB 2048x2600 00003-738036621.png)
(5.47 MB 2048x2600 00005-3826543207.png)
>>2127 Copying this from a 4chan post. Whenever I have info that's about improving AI gens, I'll just reply to this post to make searching for informative posts more easy in the future. This is on the topic of Flux and making quality backgrounds. >Do you have any version of Flux you'd recommend then? <if you feel like seperate background/character gens is too much, resharpen extension + the default forge FreeU integrated is a good way to get a lot more detail/better lighting on your backgrounds without any extra overhead or gen time like from loras or high steps (though they'll probably still look more nonsensical than flux no matter what you do, have had occasional luck with canny to improve them a bit further, but not enough to make it part of my regular workflow) <resharpen i use at 0.9 & 1-cos for both gen/hrfix, freeU i use sdxl preset with start at 0.6 and end at 0.8 <if there's any downsides to these settings i haven't noticed them Supposedly Flux_D was the version to use.
>>2147 i dont like landwhales, but that face, and those huge ears are 10/10
(1.09 MB 2400x1760 1744539755227302.jpg)
(502.81 KB 1727x935 1744549919457422.png)
>>2127 Another advice post from 4chan. This time going over inpainting and photoshopping a character into a scene. <Start with a basic sketch and high denoise, then fix different parts with a lower denoise. Do a bunch of gens and combine the best parts in photoshop >How do you blend it into the scene properly (5->6)? <1. Use Levels to decrease amount of black and make it less contrast <2. Add layer masked with the dude's shape, filled with the ambient purple color (eyedropped from the couch), mix mode Hard Light around 60% <3. Add slight hightlight on the face, hands and knee with the tv screen color, mix mode idr soft light or colore dodge.
(7.27 MB 2048x2600 00170-3440330944.png)
(7.13 MB 2048x2600 00176-3252475959.png)
(6.13 MB 2048x2600 00181-876125993.png)
(7.61 MB 2048x2600 00186-3756027530.png)
I recently learned of the prompt "gigantic ears". It works very well. >>2149 So you just want a thinner version then? Sure. I was going to go back to that gen anyway when I came up with a good outfit/theme.
(7.68 MB 2048x2600 00072-3834610304.png)
(8.01 MB 2048x2600 00084-1509757582.png)
(7.62 MB 2048x2600 00081-1509757579.png)
The lack of inpainting on the hands and face is more noticable here. They don't look as slick as they did in here. >>2154 That aside, trying to see what you can do with a crystal ball. I think the angles/poses could have been more varied in these.
(7.16 MB 2048x2600 00173-3440330947.png)
(6.93 MB 2048x2600 00187-3756027531.png)
(7.61 MB 2048x2600 00194-406279529.png)
(6.66 MB 2048x2600 00191-406279526.png)
Large breasts version.
(7.32 MB 2048x2600 00202-2736236568.png)
(8.01 MB 2048x2600 00211-1374866807.png)
(7.46 MB 2048x2600 00209-2810942042.png)
Necromancer class.
(7.86 MB 2048x2600 00207-2810942040.png)
(7.79 MB 2048x2600 00205-2810942038.png)
(7.61 MB 2048x2600 00210-1374866806.png)
Very fluffy tail.
(8.36 MB 2048x2600 00085-1768565359.png)
(7.99 MB 2048x2600 00086-1768565360.png)
(8.37 MB 2048x2600 00089-1768565363.png)
Someone finally made a good Vanillaware lora. I'll make good use of it.
(7.82 MB 2048x2600 00093-1054204736.png)
(7.87 MB 2048x2600 00096-3923472432.png)
(8.54 MB 2048x2600 00094-1054204737.png)
The white fur variant. I need to make a version that includes both versions of the girl.
(7.59 MB 2048x2600 00116-4248468082.png)
(6.78 MB 2048x2600 00119-3836788560.png)
(8.03 MB 2048x2600 00092-1054204735.png)
I made her skinny to get some more interesting body types.
(9.99 MB 2712x2712 00137-3813170106.png)
(10.47 MB 2712x2712 00147-978965844.png)
Curly hair and Panty & Stocking style. I didn't expect it to make the style look like that.
(9.11 MB 2712x2712 00144-978965841.png)
(8.08 MB 2712x2712 00154-2318941678.png)
(9.40 MB 2712x2712 00145-978965842.png)
That's all for this batch. Came out much better than expected. I love the exaggerated proportions you can get when you make girls skinny.
>>2157 woah she lost weight! i like the middle one the most out of all of them. i decided to try to lazily remake her for css using her snout shape from the image https://gamebanana.com/mods/588458
(5.97 MB 2048x2712 00275-665429487.png)
(5.40 MB 2048x2712 00285-2274275321.png)
(5.25 MB 2712x2048 00297-496988838.png)
(5.55 MB 2712x2048 00302-2718266605.png)
(5.78 MB 2712x2048 00293-496988834.png)
>>2165 Cute skin anon. Also a lot of people seemed to like that same gen on CivitAI.
(6.45 MB 2712x2048 00310-2084647499.png)
(6.51 MB 2712x2048 00316-1840827336.png)
(6.77 MB 2712x2048 00314-1840827334.png)
(6.83 MB 2712x2048 00305-283483690.png)
Storm dress
>>2125 Appreciate these tutorials though mind if i ask something more specific? how exactly do you generate younger kemoner girls instead of fluffier mature adults? similar to pic related do you just add words like "small,childlike,chibi,young,innocent,wholesome,schoolgirl" and test the prompt results afterwards? I saw some anons earlier generate kemonololi art in the previous thread, figured this would be a good starting place to try something new without fearing cucked tumblrina/plebbit censorship instantly banning me on sight. >>2147 >>2150 Hehe taking notes of these thanks! also ill get my gaming rig fixed in may. >>2128 >>2166 Would sexo Furbros ive just i wanted to interact with my waifu in peace, without being labelled a pedo for not making them look and act "aged" enough, that's really all honestly. Now technical matters aside lets say for a minute, if tomorrow i were to signup to ChatGPT or c.Ai then purchase a premium account using bank issued credit card and use it to "solicit services with characters" in a way TOS deems inappropriate (Eg roleplaying with a minor) do i get a platform wide suspension and that's simply it? Can feds actually bother to bust down my door without warning for "underage V&" Are the rumors even real or is it just some gay ass 4cuck scare meme? has anybody ever had this happen IRL how does one prevent this?
(149.27 KB 352x707 chrishansen.png)
>>2169 >Can feds actually bother to bust down my door without warning for "underage V&" Are the rumors even real or is it just some gay ass 4cuck scare meme? has anybody ever had this happen IRL how does one prevent this? i have never heard of this happening. the closest charge i can think of is soliciting a minor, but you cant solicit a minor by talking to a chat bot. you could also try not being a pedophile
(7.19 MB 2048x2800 00020-1025306675.png)
(7.23 MB 2048x2800 00025-4171196160.png)
(8.31 MB 2048x2800 00014-983682264.png)
>>2169 Post an image via catbox or Pixeldrain. Then I can see the metadata/prompt info and give you advice from there.

[ Return / Catalog / Top ]
Delete
Report