Features and Appeal of the Latest Stable Diffusion Models
Ready to Use! Summary of Effective Prompts and Recommended Settings
Hello, this is kojirom.
The world of AI image creation is overflowing with new discoveries.
Appealing new models are emerging one after another, and my disk is already full.
But this wave doesn’t seem to be stopping anytime soon.
I hope the models I’m introducing today will bring a new inspiration to your creative activities.
Animated Hard Hentai – v2.0 | Stable Diffusion Checkpoint | Civitai
[Author = Ilandis]
AnimLegit – legit-1.0
AnimLegit – legit-1.0 | Stable Diffusion Checkpoint | Civitai
[Author = SADcat]
CheckpointMirorMix – v1.0
CheckpointMirorMix – v1.0 | Stable Diffusion Checkpoint | Civitai
[Author = MaoMiror]
CLZK3 – R
CLZK3 – R | Stable Diffusion Checkpoint | Civitai
[Author = PickleJuice]
HimawariMix – v10.0
HimawariMix – v10.0 | Stable Diffusion Checkpoint | Civitai
[Author = min__san]
HotaruBreed RealMix – v4.0
HotaruBreed RealMix – v4.0 | Stable Diffusion Checkpoint | Civitai
[Author = JujoHotaru]
KayWaii – v8.5
KayWaii – v8.5 | Stable Diffusion Checkpoint | Civitai
[Author = kayfahaarukku]
Milky Wonderland – v3.0
Milky Wonderland – v3.0 | Stable Diffusion Checkpoint | Civitai
[Author = technobyte]
nidea – v1.0
nidea – v1.0 | Stable Diffusion Checkpoint | Civitai
[Author = 0922]
RE-SONATE – ORIGINAL
RE-SONATE – ORIGINAL | Stable Diffusion Checkpoint | Civitai
[Author = AI_AIDIA]
StilldopeMix – v1.0
StilldopeMix – v1.0 | Stable Diffusion Checkpoint | Civitai
[Author = stilldope]
Animated Hard Hentai
Use prompts like exposure blend, medium shot, bokeh, high contrast, cinematic tones, muted colors, and low saturation for image generation, adjusting values for hdr, soothing tones, hyperdetail, and noir to achieve desired effects. For optimal results, enable the Hires upscaler with Afterdetail for face redrawing, use DPM++ 2M Karras or DPM++ SDE Karras as the sampler, and set steps between 20-40. Use 4x-UltraSharp R-ESRGAN 4x+Anime6B for Hires upscaling at 2x, with Hires steps of 15-40, denoising strength between 0.2-0.6, CFG scale from 5-14, and skip 2 clips.
(Site Summary)
“This is an advertisement.”
FANZA Illustration Collection (R18)
Exclusively available on FANZA!
My CG illustration collection for adults, created by me, kojirom, is now on sale to rave reviews.
Please check it out!
AnimLegit
(Site Summary)
- Hello! After reviewing my first model, I noticed its limitations in handling certain painting tasks. I’ve now developed a new and improved AnimLegit model that significantly enhances pupil drawing and overall style, making your art creation process more enjoyable!
CheckpointMirorMix
(Site Summary)
- Hello, this model combines various models including goofballMix_v2, CheckpointYesMix_v35, and others. For my images, I use positive prompts like ‘anime’, ‘1girls’, ’20year old’, and ‘perfect hands’, and avoid negatives like ‘text’, ‘watermark’, and ‘bad quality’. My Loras settings include ‘add detail:1.0’ and ‘body drawing control:0.5’. The images are created using Hires Fix Scale with 2xAniScale for upscaling, and Ultimate SD Upscaler and FaceDetailer from ImpacktPack for finishing. The process involves steps of 100 and 10, with a cfg of 8, using euler_ancestral sampler and karras scheduler.
〈This is an advertisement〉
Kojirom’s Adult CG Illustration Collection Now On Sale!
An adult CG illustration collection by kojirom is now available on FANZA.
Starting from just $1.0
It’s highly praised for its delicate touch and unique worldview.
Don’t miss this opportunity to check it out.”
CLZK3
(Site Summary)
- 半分はCLZK2、半分はSimiKawAmaの0.5とkawaiiMonke_v100の0.3とmergedAmateurs_mixedAmateursの0.2を合わせたものをさらに0.35かけたもの）の33.4％
- BamAuMei（brickandmortarmix_v30、aurora_v10、meinahentai_v4をそれぞれ33.3％ずつ合わせたもの）の15％
- TestTraidDiff（easyfluff_v10PrereleaseとCLZK2Sからfluffyrock_e184VpredE157を引いた差分の違いを1_trainDifferenceで掛け合わせたもの）の33.3％
HimawariMix
Models with built-in VAE with strong expression of backgrounds and details merging various models
(Site Summary)
〈AD〉
Exclusive Sale at FANZA!
My CG Illustration Collection (for adults) by kojirom is now available and highly acclaimed.
Please check it out!
HotaruBreed RealMix
The ‘HotaruBreed’ series, version 4.0, focuses on enhancing ‘kawaii’ or cuteness, with improved color saturation and shadow intensity. These models are akin to flower breeding for aesthetic enhancement. HotaruBreed features include a variety of trained LoRA’s for style adjustments, detailed backgrounds, and NSFW support. The ‘RealMix’ special edition offers a more 2.5d-like style. Explore various HotaruBreed models on my page.
Recommended settings include CFG Scale of 7-10, Clip skip 1-2, Samplers like DPM++ 2M/3M SDE Karras, 30-50 sampling steps, and resolutions tested on various sizes. Use HiRes Fix with SwinIR_4x or R-ESRGAN 4x+, denoise strength 0.3-0.6, and negative embeddings like EasyNegative. For VAE, use ClearVAE 1.0 for versions without VAE. Quality prompts like ‘insanely detailed’ and ‘ultra cute girl’ are suggested. Check out more LoRA’s on CivitAI models list or HuggingFace repository.
(Site Summary)
KayWaii
Maintaining the KayWaii stable diffusion model is resource-intensive, so support via Ko-fi is appreciated. The model, focused on anime-style girls with thick linearts, is always free. It uses merging methods similar to AingDiffusion, another anime model I created. For best results, use the included VAE, EasyNegative for negative prompts (avoid EasyNegativeV2), and “DPM++ SDE Karras” sampler for versions 2.1 and up. High-resolution images require Hi-res fix with SwinIR 4x or 4x-AnimeSharp upscaler. If VAE causes issues, set it to ‘None’.
FAQ highlights:
- Frequent updates are due to my personal use and continuous improvement of the models.
- For better image generation, adjust prompts, sampling methods, steps, or CFG scale.
- AingDiffusion and KayWaii differ in goals and styles; AingDiffusion has a pastel anime aesthetic, while KayWaii focuses on cute girls with thick linearts.
Experiment and enjoy the models!
(Site Summary)
Milky Wonderland
Milky Wonderland focuses on creating images that embody cuteness and wonder, featuring charming characters and mesmerizing sceneries. The v3.0 update enhances details, backgrounds, and scenery. This model is available online on TensorArt. Users must adhere to CivitAI’s Terms of Service when generating images with this model. Special thanks to Kybalico for their contribution.
(Site Summary)
nidea
Just another pointless merge that no one will ever use.
It would be nice if you could show me something made by this model.
(Site Summary)
RE-SONATE
RE:COLLECTION introduces ‘RE:SONATE’, a 2D model for creating stylish characters and detailed fashion, with versatile color tones and high-quality backgrounds. It’s part of a series including models like RE:QUIEM and RE:LIVE. The original RE:SONATE is 3.97GB, with a lighter version coming soon. Recommended settings include 30 sampling steps with DPM++ 2M Karras, CFG scale 7, and specific denoising and upscaling parameters. Positive prompts like ‘best quality’ and negative prompts like ‘badhandv4’ are advised.
(Site Summary)
StilldopeMix
Ayo, this merged model combines various others and handles both SFW/NSFW content well. To avoid stains, use a VAE with Euler a, 25 steps, and a CFG Scale of 5. Enjoy using it and share your great results!
(Site Summary)
All Illustrations Are Commercially Usable
Some illustrations may have incomplete details such as hands and fingers. We kindly ask for your understanding.
When using these images, please feel free to crop or modify them as needed.
We encourage you to fully express your creativity and use these illustrations with a free and imaginative approach.
The illustrations featured on this site are available for commercial use.
