Kavka Mix (Version 1.6)
v1.6
Kavka
An update to go with 1.5 and above.
Example prompt of a city rat:
uploaded on e621, anthro, nsfw, explicit, explict, (rat, skaven:1), high res, (detailed realistic image:1), (detailed eyes, beautiful expressive eyes:1), impasto impressionism, insane details, soft, (hyper realistic fur:1), (detailed fur:1), pupils, (fur covered body), (male:1), (slightly chubby:1), (rat features, skaven features:1), (expressive face, detailed face:1), (realistic:1), (photorealistic, photo-realistic:1), nipples, belly button, full color, (3d:1), (highly detailed:1.2), masterpiece, 8k uhd, (suspicious expression:1), (back alley background), (full-length portrait), pants, open jacket, peeking around corner, (athletic:0.8), (two tone hair, anime hair:1), wielding pistol, dark fantasy art
Negative prompt: (worst quality, low quality:1.4), ugly, ((blurry)), bad anatomy, disfigured, extra limb, bad-artist, (EasyNegative:0.5), boring_e621, (muscle, muscular, overweight:1), (breasts:1.4), alien, bara, simple background, wings
Steps: 25, Sampler: DPM++ 2M Karras, CFG scale: 7, Seed: 609369693, Size: 512x768, Model hash: 2718b4565e, Model: KavkaMixV1.6T2, Denoising strength: 0.7, Hires upscale: 1.5, Hires upscaler: Latent
You can use tags like these in the prompts or the negatives to influence overall style: uploaded on e621, anthro, alien, bara, dark fantasy art, kemono
A curated blend of herbs and spices. A touch of fantasy, a whole lot of fur, and a wide range of themes. It can do human, furries, hybrid monstrosities, and more!
Want twinks, muscle hunks, or chunky dudes? It can do a wide range of body types. I also do believe it can do other genders just as well.
Want some more cutesy style? Throw kemono into the prompts!
A bit of darkness? "dark fantasy art" will surely help!
Want to start playing around with porn? nsfw, explicit, explict for prompts. For furry, include anthro, furry, uploaded on e621, and include (by artist1, artist2, etc). Take a look at my example pictures' PNG info for more examples! Prompts are mainly e621 based, and I find no underscore is required, but some will use apostrophes and hyphens!
My preferred settings: DPM++2M Karras, 20-50 steps, CFG 6-10, 512-768x512-768, Upscaled 1.5 with Restore Faces on.
I use the WD-KL-F8-Anime2 VAE. There's none baked in, and seems pretty flexible with that, as well as KL-F8-Anime2 and anything-4.0. It depends on my mood.
This checkpoint also works very well with all the Loras I had tested it out with. It certainly loves the 3d render ones. Best strengths range from 0.3 -0.6 for stylistic and thematic ones, and character ones around 0.8.
Negative prompts: (worst quality, low quality:1.4), ugly, ((blurry)), bad anatomy...I also suggest EasyNegative:0.5, bad-artist, and boring_e621.
Happy generating!
You can support by buying a coffee ☕️ here —
What is Kavka Mix (Version 1.6)?
Kavka Mix (Version 1.6) is a highly specialized Image generation AI Model of type Safetensors / Checkpoint AI Model created by AI community user Kavka. Derived from the powerful Stable Diffusion (SD 1.5) model, Kavka Mix (Version 1.6) has undergone an extensive fine-tuning process, leveraging the power of a dataset consisting of images generated by other AI models or user-contributed data. This fine-tuning process ensures that Kavka Mix (Version 1.6) is capable of generating images that are highly relevant to the specific use-cases it was designed for, such as anime, sexy, porn.
With a rating of 5 and over 12 ratings, Kavka Mix (Version 1.6) is a popular choice among users for generating high-quality images from text prompts.
Can I download Kavka Mix (Version 1.6)?
Yes! You can download the latest version of Kavka Mix (Version 1.6) from here.
How to use Kavka Mix (Version 1.6)?
To use Kavka Mix (Version 1.6), download the model checkpoint file and set up an UI for running Stable Diffusion models (for example, AUTOMATIC1111). Then, provide the model with a detailed text prompt to generate an image. Experiment with different prompts and settings to achieve the desired results. If this sounds a bit complicated, check out our initial guide to Stable Diffusion – it might be of help. And if you really want to dive deep into AI image generation and understand how set up AUTOMATIC1111 to use Safetensors / Checkpoint AI Models like Kavka Mix (Version 1.6), check out our crash course in AI image generation.
Popularity
Info
Latest version (v1.6): 2 Files
About this version: v1.6
More realism, and a lot more added details. Now you can get away with simple prompts for a variety of things, like sweet kick ass scenery, or more complicated for detailed characters. Please look at the prompts contained within the pics to get a rough idea on how the prompts work. You can go as complicated as mine, or you might find that fewer will work.
Again some pics will be re-generated from previous versions, but I'll be adding a lot more newer stuff to show other capabilities.
A good beginning set of prompts for characters go along like this: high res, (detailed realistic image:1), (detailed eyes, beautiful expressive eyes:1), impasto impressionism, insane details, pupils, (expressive face, detailed face), (realistic:1), (photorealistic, photo-realistic:1), (full color:1), cinematic lighting
Negative: (worst quality, low quality:1.4), ugly, ((blurry)), bad-artist, (EasyNegative:0.5), boring_e621
Trigger words to change styles include: science fiction art, dark fantasy art, kemono, bara, anthro, alien, uploaded on e621, fantasy, apocalyptic, etc.
Also included for overall results: pencil sketch, ink art, watercolor art, disney style art, cartoon art, photograph negative, and so much more. Try to think outside the box with some!
Remove the negatives if you're aiming for sketch or more simplistic styles, or if you want to degrade the quality.
Look at the pictures to see their full prompts and steps. 20-25 steps, 6-8 CFG is good, though I've seen other people use 40-100 steps with higher CFG like 12-15. As always, it can really depend on your graphics card and drivers.
7 Versions
Go ahead and upload yours!
