If you enjoy my contribution to this community, feel free to buy me a coffee: the more caffeine I drink the more models I can create 😅
Comparison gallery here: Fv6-Fv5 and NSFW illustration comparison between Gv1-Cv6-Fv2.
Current SOTA model in my experiment:
SDXL Base model: Gv4 is the most balanced model allowing both realistic and styled NSFW and SFW images. Better aesthetic than Fv6 but less photorealistic.
SDXL photorealist (SFW and NSFW) model: Fv6 is the way to go for hyper-realism including realistic NSFW images but it mostly lacks the styling capabilities of Gv4.
SDXL Illustration : Gv4 (SFW and NSFW). Cv6 is however still worth a try if you are not into NSFW images.
Flux model: Flux1-A1
Image Generation Settings for SDXL models
DPM++ 2/3M SDE / Karras or Exponential are always a good bet with 25+ steps and CFG around 5-7. But DPM++ SDE / Karras with less steps (e.g. 12) and higher CFG (8-11) is worth a try as is Euler Ancestral / Normal for a less detailed result which may also be interesting with a model like Fv6 to get something less "noisy".
The default CLIP Skip of 2 is also a good bet, but using 1 or 3-4 is also worth trying: 1 push more towards prompt adherence and 3-4 give sometimes a better result than the default focusing more on the "concepts".
Image Generation Settings for Flux models
My preferred settings are DPM++ 2M / beta or sgm_uniform or DDEIS / normal for the sampler / scheduler, beta giving a bolder stronger image. For a more subtle image, Euler / simple or beta seems a good bet.
CFG seems to have a huge impact on the final image and be very sensitive even to small variations.
For photos, CFG should remain low (1.5-2.5) to avoid plastic skin.
For fine art and illustration it is more complicated because it depends on the medium. For "rough" styles (painting, watercolours etc.), CFG should stay quite low in the 1.5-2.5 range but for anime or comic style, CFG needs often to be pushed further to achieve the desired style (3-6 or more).
If the image is messy/malformed or blurred, it is often because the CFG/steps are inappropriate for this image, but it is not always easy to know whether CFG/steps must be increased or decreased (at least to me 😊).
There is for sure a lot to learn concerning Flux behaviour which is quite different than SDXL and we will need to adapt.
Workflow
All my images are uploaded with embedded ComfyUI workflow which is alas incompatible with CivitAI processing and most often prompt cannot be retrieved. You can however download the original PNG image with the workflow included by clicking on the "DOWNLOAD" icon in the image viewer.
Why I try to mostly publish images straight from my model with maybe a bit of a Lora (mine or some detail enhancer), I also sometimes now use Controlnet to get better more detailed compositions more easily: in this case the source image is obviously not in the workflow, but I guess you can still use the image I publish as a source if you want to make a variation :)
Past
Starting from E serie, models are evolving sometimes by merging with other models (thanks to other contributors!), but mostly via training on my own dataset: a modest dataset (~2000 images currently), but I try to somehow compensate with quality and originality.
Starting with Fv1, I have included many synthetic images I created using previous versions: playing hard with prompt and retouching when necessary the result in Photoshop in order to have a dataset which contains many original images.
The core idea behind this model was to create a versatile tool by merging some of the best existing models which fits my personal taste (photography and fantasy art to make it simple). My primary goals were:
Photorealism: The ability to produce stunningly realistic images of both people and objects/nature.
Flexibility: The ability to create highly stylized images, allowing for artistic expression through various styles and combinations of artists. I am from a older generation and comes from Europe so "style" does not mean for me "Japanese kawaii aime with boobs" or "DC Comics cartoon with lots of superheroes and voluptuous blonde babes" but more from a univers of Frank Frazetta, Milo Manara, Boris Vallejo, H.R.Giger, Wojtek Siudmak and such fantasy art masters: there are boobs involved for sure, but the style is somewhat different :P
As I doesn't like being limited in my exploration of the human body, the idea is also to have a fairly capable NSFW model. However, dur to the nature of available training images in the data sets, NSFW often comes with a strong bias toward either porn photos or porn Japanese anime and impact the flexibility (typically as soon as you use the word "sexy" in your prompt you need to weight in the style). This point is thus NOT the priority for the base model but is pushed forward in the NSFW model.
Description
Ev1 was a new starting point a bit rough and it didn't last long 😅.
This evolution Ev2 improves the overall quality of images a lot.
Based on Ev1, only Loras were merged: add_details and xl_more_art-full_v1 as well as two personal Loras (araminta-s-glamourphotography as well as an unreleased secret one ^-^), which means that for the first time this model is no longer a simple merge but actually contains parts which are trained on my own dataset.
FAQ
Comments (5)
This model is so good, its almost half way to a pony model. Its compatible with many pony loras, I recommend envy zoom slider for NSFW content. Ev2 lacks a lot of styles that Cv6 has. But that is easily solved with img2img styles. If you want to also be able to get illustrative results of various kinds I'd recommend Cv6. Everything that is realistic Ev2 is very good at. great eyes, good faces, excellent posing. decent hands :) since it gets basicly 75-90% of the way and handles Pony models, you can mix in either XL models or Pony models for advanced poses on very low strength and get the rest of the way without sacrificing your face or style. Penises are not always great though, huge part to resolve that is using zoom slider. maybe with a penis lora or smth it can get easier?
Agree : it is very good at NSFW but Cv6 is a bit better or at least easier for styling. I will try to work on my own Loras in the futures to try to push the models in a more controlled direction. Thanks for your feedback!
@aramintastudio great models all of them though. only downside is that it hogs so much diskspace for me now to have 3 of them :D cause I will never give up on cv3, and cv6 is like cv3 but cleaner, And now Ev2 is like the others but probably better at realistic. better at NSFW yeah, havent quite figured out what I like more in terms of SFW realism think its Ev2 but need a lot more time with it. I missed your comparison link that explains nicely the differences :D I'm gonna have to try some style img2img exports from Cv3/Cv6 into Ev2 see if I can get some cool cartoonish stylish nudes that way.
@GreatHunt It is indeed the problem I tried to solve (at lease for myself :) : all good models (and there are many) are good at something and not as good at other things. I try to create if possible 1 but maybe 2-3 models which are best for ME and hopefully some will please to other people here.
I personally try to completely forget my old models and concentrate on the 1-3 SOTA (Ev2, Cv6 and maybe Bv4) even if I indeed know than old one were very good as specific things.
@aramintastudio I think cv3 is great as is even with a few faults, if I want more anime I can look for other models. cv6 is great as is, its like perfect balance, it can remain like that even if expanding or taking things from it can be valuable. I don't mind the realistic direction of Ev2, its great,
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.