I try to pin images that stand out / i personally like every 1-3 weeks
So if i unpin yours doesnt mean that i dont like it anymore just trying to keep a fresh rotation of images without completly filling the start page with pins <3
Join the Discord for questions / sharing pictures :) : https://discord.gg/gcJqAKQ5Af
If you enjoy my models and want to support me, you can now support me on kofi
(Since my patreon got disabled like a year ago & buy me a coffee banned me 4.2026 )
🛠️ Recommended Settings:
Resolution:
best results with832x1216//960x1440//1024x1024//1024x1536CFG Scale:
5–8Sampling Method:
Euler aorDPM++ 2M KarrasSteps:
20–50
Description
FAQ
Comments (22)
what are the main changes for v8? i dont see any crazy difference from v7
V8 isn’t a huge stylistic jump from V7. It’s more polished overall, with better refinement and noticeably less color saturation. Subtle improvements rather than dramatic visual changes. :) forgot to put it in changes my bad
@Desync how up to date is the checkpoint in regards to knowledge of characters?
I find it wee strange you got to v5 and chose to pivot away from semi-realism? We have a great, many, hundreds of anime IL models yet the only ones that came close to something like 2DN was for Pony, and Deliberate v2 was for SDXL, are yours and Wicked.
hi i was just doin what most people seemed to like more, we already voted on discord tho to make a new model that goes more in direction anime and keep this in the other style :)
to be fair v3 was by far the most used one that was the first step more towards anime style
@Desync Ah, understandable. Give people what they want and all that. Semi-realism does seem a bit niche, overall. But I've always preferred it because its capable of fine details, while the realistic models are all trained on medium, which crushes things like silk, nylon, and glass, oddly enough. Either way, best of luck whichever path you choose next!
I'd love to see a semi-realistic 'branch' of this model. I really liked where you were going with v5. (The new versions are great, but not the semi-real, as OP noted)
@ttb2112Â v9 prolly comming tomorrow and we going back to semirealism with it :)
@Desync AWESOME! Thank you!
Nice, thank you. I will try some older prompts, to compare with others models. But I am more than satisfied with outcome already.
Prompt example?
all images i posted have the prompts avaiable :)
16gb VRAM minimum required?
no ?
I'm using 8gb Vram, works fine.
It's SDXL based, so no and never. Unless someone for some reason took SDXL and scaled it up somehow and made a entirely new model, with only thing being the same is architecture which would be incredibly stupid since there is way better model architectures now
pipe.enable_attention_slicing()
pipe.enable_vae_tiling()
those are your best friends
Worked even on 1050GTX 2Gb at least in Comfy. Slow, but worked without problem up to 2mpx(no upscale) and up to 1.5x more with upscale. Just manual override torch to 2.8.0+cu128. It was more RAM issue in my case, but when i set 48gb RAM it isn't a problem now.
no. mine works fine using Draw Things on my iPhone 13.
Hey, I’m building a prompt workflow tool called MorpBase at morpbase.com. It’s focused on structured prompt building and iterative image-creation workflows, so I’m always interested in how people use models like this in practice. If you’re a very active generator and would be open to testing it, feel free to DM me.
Great, but suffers from tumblr nose syndrome.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.