Find all my models, scripts, and beta versions on my Ko-Fi.
☄️ IntoRealism
IntoRealism is a model designed for high-fidelity realism. From lifelike portraits to nuanced lighting and natural textures, this model brings your prompts to life with stunning detail, without sacrificing flexibility or creative control.
📢 Join The Community
We're building a friendly space to chat, share creations, and get support.
👉 Click here to join the Discord
You can visit my Ko-fi page if you appreciate my work for a small donation, orders, etc.
⚡ DMD2 Usage
Sampling method: LCM / Exponential or Karras for V3.5
Sampling steps: 8 to 12
Resolution: 896x1152 / 832x1216
CFG: 1Upscaler
4x-UltraSharp or others
Denoising strength : 0.3
Hires CFG Scale : 1Adetailer
Adetailer model: face_yolov9c.pt
Inpaint denoising strength : 0.25 or 0.30SDXL Usage
Sampling method: DPM++ 2M SDE Karras
Sampling steps: > 25
Resolution: 896x1152 / 832x1216
CFG: 4-5Upscaler
4x_NMKD-Siax_200k or other
Hires steps : 4
Denoising strength : 0.3
Hires CFG Scale : 4Adetailer
Adetailer model: face_yolov9c.pt
Inpaint denoising strength : 0.4 (long view)ZIT Usage
Sampling method: Euler / Beta
Sampling steps: 6/10
Resolution: 896x1152 / 832x1216
CFG: 1Description
V2 is more realistic than the first version. We might lose a bit of the NSFW vibe, but not too much. More LoRAs have been merged in.
FAQ
Comments (24)
Wow ! Nice work !
Nice model, can you upload one version withou dm2?
Maybe in June.
This is Just too good! Try using any lighting prompt and it creates amazing images. This is a GOAT model!
Thanks !
Out ouf the box the lighting needs some advice - I'm with you. I'm using the great gonazLomo checkpoint a lot and gave this here a try. Very nice I have to say.
V2.1 has a higher % of 3 finger hands compared to earlier versions. Also, in some instances, the skin pores are overwhelming. However, a negative helps in some cases. The prompt adherence is extremely good. Also, the way this model gives new poses for every run of the prompt is simply too good!
@zepled23 I will try to solve this problem in the next version.
The V3 series is in a league of its own. First look at 3.2 - just wow.
I must say, V3.2 is a whole another level. Expressions are so good! Landscapes are amazing!
The best SDXL model I've every used so far.
I think 4 steps is enough, I detected no quality difference comparing to 8. Anyways, this model is extremely good!
The details are slightly better at 8 steps, but 4 steps still works great.
I normally use 12 steps with Dmd and i can tell the difference with only 8 steps.
V2.1 is an overall improvement on the realism side. I'm currently working on V3, which will likely be much more SFW, featuring landscapes, themed content (cinema, amateur), and an even higher level of realism. I'm always open to feedback and ideas to make this model even better! :)
More boobies thanks.
Oh Hi.
I'll start for the cons: Lightning falls in the flat side. It's like colors are always hidden behind a grey cast so there's little volume to the pictures. I think more contrast might look better (don't overdo it). It's like it needs to expand a little bit the dynamic range.
Pros: Colors are not saturated. I've noticed this trend in a lot of checkpoints and it drives me nuts. There's no easy fix to oversaturation. But this checkpoint has a more neutral palette. Truth to be told, It's more on the cold side of the spectrum, but it can be easily fixed. So that's a good starting point.
All in all, It's a very promising series I like the skin detail and overall aesthetics (that doesn't involve colors). Anatomy seems strong too.
There is still room for improvement, but I'm interested in the future checkpoints. It's a great start, all in all.
Keep up the great work!
Thanks for your feedback!
I pretty much agree with the “desaturated” aspect. I’ll work on that.
As for the “brightness” part, I feel like my model actually stands out a bit from the others. It’s not perfect yet, of course, but it’s far from being as flashy as 90% of the models out there. If you’re using Hires Fix, you might also want to check which upscaler you’re using, as it can make a difference.
I do think there’s room for improvement, and I’ll focus on that over the next few days to make V3 even more refined and polished.
Thanks again for your feedback. I hope the next version will better match your expectations!
@enzino just a thing. I've said contrast not brightness (i'm not sure if brightness is enough or not unless contrast is corrected first). No i don't want flashy checkpoints either. Just normal colors. I'm a firm believer a base model should strive for a neutral color palette (as much as i possible) and avoid color biases. Yes, I know what are the VAE limitations of sdxl -i've been struggling agaisnt it for months- but still... it can do better colors. Mm... Talking about this is such a coincidence!, I don't know if you use loras, but i used this one the other day and i really liked it the outputs. I hope it helps someone. After several days looking for color tweaking loras, this was the one that gave me the best results: https://civitai.com/models/1363766/sdxl-color-tweaker-slider-color-temperature-saturation-brightness-contrast-or-sdxl
@ElectricDreams thanks for sharing the LoRA, I’ll definitely check it out and see how it looks. Appreciate your help.
@enzino you're welcome. It's the only one i've found that gives you control over saturation-contrast-temperature-brigness. It's actuall 4 loras. It helped greatly with my outputs. A lot of people use comfy nodes to tweak color settings in the checkpoints but i don't want the hassle. I'm too lazy for that.
@ElectricDreams I found a way to boost the contrast (not with this lora). I might release a V2.1 soon.
@enzino By any chance, did you merge a Pony checkpoint in "intorealism"? that absolutely would explain the noise and the lighting. Just out of curiosity, you don't have to reply if you don't want to.
I like DMD2 so I gave this a try and I love the results - just some corrections in the lighting and it's great. You did a great job - thanks!