I believe this model to be very good at photorealism although it does have some artistic capabilities as well. Here are a few tricks to using the SDXL versions of GonzaLomo to its fullest potential:
Clip Skip is your friend. I recommend trying a prompt with both 1 and 2 values. 1 can make images that are more creative, artistic, retro or grungy whereas 2 is better at photorealism and the, um, physical attractiveness of the subject(s).
Generate at 1024x1536 or higher using the PatchModelAddDownscale node in ComfyUI. This will create a much more detailed image that can then be upscaled further, resulting in incredible photorealistic detail. Using that node, I can do an initial generation as high as 1280x1920.
You can play around with different sampler settings but I find that the absolute best results are with LCM/Karras, 12 steps and CFG 1.0.
Selective usage of the Self Attention Guidance node can sometimes give amazing results. I always have it in the pipeline and usually bypassed but turn it on when I think an image has potential but just needs a little more oomph like color saturation or contrast.
I get really interesting/creative results by generating with the V1.1 Semi-Real model and then resampling that image with the V2.0 Unity model at around 0.60 denoise. You really need to play with the denoise to find the right value that tips the image from cartoonish to photorealism but you'll be amazed at what's possible when you do. Note that V1.1 Semi-Real requires Clip Skip 2+ and will not work with 1 at all.
And lastly, always be upscaling! More pixels equals more detail equals more photorealism.
Note: I only use ComfyUI so I apologize in advance if you're having trouble with my model in some other tool and I can't help.
Current Models
V2.0 Unity
Essentially, this is V1.1 Real with a bit of Siren Forge, Cyberrealistic and Lustify added to the mix. The base is still Gonzales and LomoXL.
I called this version Unity because I originally wanted to have separate models for photorealism and artistic styles. However, I realized that I could get both in one model and created V1.1 Artsy for that purpose but it was a failure. This is my second attempt at that combination and I think/hope it's a success.
At the very least, I don't want to degrade the photorealism capabilities and just want to add more surreal, artistic or fantastical styles as an option.
The recommended settings haven't changed since V1.1 Real:
Sampler: LCM (also dpmpp2a and ddim can provide interesting results)
Scheduler: Karras, Exponential or Beta
Steps: 8-12
CFG: 1.0-1.3
Clip Skip: 1-5
I normally keep my settings at LCM/Karras with a CFG of 1.0 but I tend to vary the clip skip between 1 and 5 because those can change the image pretty drastically. Sometimes an image at clip skip 2 will be terrible but might become amazing with clip skip 1, and vice-versa.
V1.1 Semi-Real
I'm doing a little experimenting with non-photorealistic models. This is a merge of GonzaLomo v1.1 Real and ToonComix which is a toon model that does a really good job with detailed backgrounds. The whole Comix family of models is very good so check them out.
Sampler: LCM
Scheduler: Karras, Exponential or Beta
Steps: 8-12
CFG: 1.0-1.3
Clip Skip: 2-5 (1 is not supported)
Retired Models
V1.1 Real
This is a simple 60/40 merge between two of my favorite photorealistic models - Gonzales-NSFW-PonyV1/V2-DMD v2.0 and LomoXL (with a little bit of Lustify thrown in). Many, many thanks to the makers of those excellent models.
Gonzales is a Pony model so GonzaLomo inherits all of those capabilties. LomoXL has a very strong analog feel to it which helps give an edge to Gonzales' "pony-ness".
Sampler: LCM
Scheduler: Karras, Exponential or Beta
Steps: 8-12
CFG: 1.0-1.3
Clip Skip: 1 allows greater analog effect but lessens NSFW capabilities
Clip Skip: 2-5 lessens/removes the analog effect but allows more NSFW
Try this workflow to get started, if you need it.
To be clear, I created this merge for my own use cases. It's certainly not better than Gonzales for most things. I'm just hoping to inject Gonzales with some of the gritty, analog feel that I was looking for in some cases. LomoXL also allows for a wider variation of faces, especially with clip skip 1.
Enjoy!
P.S. This is my first model upload so I would appreciate any constructive feedback or comments.
V1.1 Artsy alpha (Not very good imo, use V2.0 Unity instead)
This was an attempted merge between V1.1 Real and a few art-focused models. The main purpose of this model is to be able to produce artistic styles but it has the tendency to try to force photorealism when prompted clearly for an artistic style. This model has been abandoned in favor of V2.0 Unity which does a much better job of combining the photorealism of V1.1 Real with artistic styles.
Description
Essentially, this is V1.1 Real with a bit of Siren Forge, Cyberrealistic and Lustify added to the mix. The base is still Gonzales and LomoXL.
I called this version Unity because I originally wanted to have separate models for photorealism and artistic styles. However, I realized that I could get both in one model and created V1.1 Artsy for that purpose but it was a failure. This is my second attempt at that combination and I think/hope it's a success.
At the very least, I don't want to degrade the photorealism capabilities and just want to add more surreal, artistic or fantastical styles as an option.
The recommended settings haven't changed since V1.1 Real:
Sampler: LCM (also dpmpp2a and ddim can provide interesting results)
Scheduler: Karras, Exponential or Beta
Steps: 8-12
CFG: 1.0-1.3
Clip Skip: 1-5
I normally keep my settings at LCM/Karras with a CFG of 1.0 but I tend to vary the clip skip between 1 and 5 because those can change the image pretty drastically. Sometimes an image at clip skip 2 will be terrible but might become amazing with clip skip 1, and vice-versa.
FAQ
Comments (34)
I'm experiencing very orange and muddy-looking images when using ControlNet with your model. I've tried different depth models and downloaded several VAEs, but the issue still persists. Could you help me figure out what's going on?
Ah, I'm sorry to hear that. I haven't used ControlNet in a long time so I didn't do any testing with it. Are you using ComfyUI? If so, do you have an image with workflow you can post?
@GBRX Thanks for replying. I actually use WebUI. What I do is generate the image using ControlNet, then manually correct the colors in Photoshop before sending it back to img2img. It works, but I feel like it could be better — other models seem to handle it more naturally
@GBRX I increased the resolution to the maximum, and it seems to have fixed the issue. Congratulations on the model — it's incredible, one of the best if not the best I've used. Just a quick question: is there any chance you might create a version specifically for inpainting?
@pabollostore205 Oh, sorry that was necessary but that's great you found a good workaround for the issue. Thanks for all the feedback. I'm probably an intermediate image generator but, in all honesty, I'm a complete novice at model creation. This is my first one so I've got a lot to learn. Rest assured that as soon as I learn how to create a decent inpainting model I'll put one out. I'm actually creating these models for my own purposes but I love that others are finding them useful too so I'll try to do more.
Your model is needed on tensor.art
I have to admit that I don't know much about tensor.art. I'm not opposed to it but what's the benefit of posting the model there?
Well, first of all, more people can understand the benefits of your model...Secondly, since I'm not a creator, I don't know either what other benefits there are.
@GBRX Please ignore my immature suggestion. I just like your model very much
@onizuka918 No, I appreciate the suggestion. I just don't know much about tensor.art yet. I'll definitely check it out. Thanks!
@onizuka918 I finally got it up on tensor.art - https://tensor.art/models/869231080954158617
@GBRX Great! However, there is a small problem. I don't know why it keeps showing "The model is being deployed. It will probably take about 10 to 15 minutes to wait."
@GBRX And it can't be found in the search either。:(
@GBRX It's ready to use now. Perfect
@onizuka918 Awesome, glad to hear it! I had a number of problems getting it uploaded and visible. Now I'll have to get the v3.0 version up there.
Best XL model ever!
Wow, thanks!
@GBRX No, Thanks to you!!!!
hi, i use forge because comfy is too complex for my knowledge, i am not a creator and i am here just to have fun. Your models in my opinion are fantastic, with my rtx 3090 at the resolution you recommended i can generate in just 3 seconds a perfect image with a magnificent resolution, among all the sdxl models i have tried and for the few tests i have done your models are certainly among the best.
i hope later you will also make some flux versions.
thank you very much for your contribution to this community.
Thanks so much for the great feedback! I'm happy the model is working well for you. I only use ComfyUI so I never know whether my model works with other tools.
I'd like to do a Flux version eventually but I only have a RTX 4060 with 16GB so that might not be possible until I can get something more powerful.
@GBRX I understand how difficult it can be to train a Flux model maybe one day, if you manage to do it, I’ll be happy to try it!
As I mentioned earlier, I usually use Flux models only with Forge because I find Comfy a bit too complicated for my current skills. I just do this for fun.
I’m testing this model now and, aside from a few challenges generating images of older subjects (which often get solved by refining the prompt), I must say it’s incredibly realistic!
It’s lightning fast, and even with very few steps, it amazes me with the stunning image quality. Even without using a rescaler, the resolution and detail are insane honestly, I’ve never seen anything like it in any other SDXL model.
Simply fantastic!.
Thanks for sharing it and sorry for my "Italian-style" English!
@jazara930 Thanks again for the excellent feedback. Does Forge allow you to change the Clip Skip? I understand what you mean about generating older subjects and I think changing the Clip Skip to 1 will help with that. That's been my experience in ComfyUI.
You can get very different images - sometimes better, sometimes worse - simply by changing the Clip Skip and keeping all the other settings the same. I've used values from 1 to 12 just to see what it would produce and some of my better images have come from using Clip Skip 1.
It's worth a try if Forge allows you to change it.
Also, if you want older subjects then try using the terms "mature" or "older" in addition to giving a specific age like "50 years old". So I might use something like "mature adult woman, 50 years old, ...". That tends to work for me.
wow! This model is very good! Thanks for sharing. I'm thrilled
Awesome, thanks for the feedback!
Your v1.1 Real model is outstanding. The image gallery caught my eye, and it's brilliant with realistic images.
Thanks so much! Please try v2.0 Unity if you can. I think it works just as well with photorealism but I'd be interested to hear your feedback.
@GBRX sure, I was curious to try v2.0 as well. My honest opinion on both the v2.0 and v1.1 is that they deserve to be featured model on Civitai. I'm quite excited and glad to see the level of perfection you've brought into 'SDXL' models. Just beautiful creativity.
@torqx I appreciate the compliments, although all I did was merge a few models others created and then hoped for the best. ;) I'm just glad people find it so useful.
I love this model. Everything happens the way I envision it. I don't have to struggle against it, like with some other models, to get what I want. The quality is excellent in my eyes and there is certainly still room for improvement. Moreover, I received valuable tips from the creator of the model and he was always there for my questions. Great praise and heartfelt thanks!!!
Thanks so much! I also appreciate all the excellent images you post and the great suggestions you have to make the model even better.
This is a masterpiece! Amazing!
Thanks!