While I'm sure Devilmarkus appreciates your astute observations, I can't help but wonder if your critiques are born from genuine feedback or the green hue of envy. It's not every day that I encounter someone with such a keen eye for detail—almost as if they're trying to divert attention from their own shortcomings. But I"m sure they're thankful for the encouragement, and hope that your projects be as flawless as your judgment.

I always reserve judgement of a model based on example images alone. If I can't be bothered to run my own gens and see for myself, why should anyone care what I have to say? Run some gens, post your results, then critique. In this case, I can't say what this model might do based on the examples, they vary pretty widely. Even so, "epic fail" is harsh considering the examples, and without your own gens to back it up, this comment just doesn't hold water.

@MysteryWrecked I am not saying it doesn't have potential, but they stated the model was meant to be a photorealistic model specifically and then provided many examples, none of which are at all photorealistic. I will try the model myself and see what it can do. My comment was based on the stated purpose of the model vs the examples images that showed none of that quality. Maybe it's a great model with very poorly chosen examples. I guess we'll see. If the model is great, I will certainly say so after trying it out.

@sgqw0n8ib243 My critique is based on the stated purpose of the model vs the example images provided. Nothing else. They said it was meant to be a photorealistic model, yet zero of the examples look even close to photo quality. For me, any claims made should be backed up by evidence, which means if the purpose is photo quality, but none of the examples are photo quality, then I stand by my comment until I see evidence that changes that.

@Devilmarkus I downloaded it, but haven't got a chance to try it yet. I will shortly.

@AlBat One thing I know, is that 1.5 model is amazing. A lot of the top creators in 1.5 are still developing best methods and practices in XL at this early stage, and many of them haven't elevated to their 1.5 status in XL. Nonetheless, you can count me as a Devil worshipper ;)

@Devilmarkus I tried it out and it's pretty good. I will keep playing with it and see if I can get any truly photoreal stuff with it. I posted my first 2 in the gallery.

@Devilmarkus No, it seems OK. I haven't had time to do too much with it, but I got a couple that looked pretty good.

@AlBat Pretty sure ALBAT didnt see the pictures of Elvira. Click to the right bro.

@kpizzle Unfortunately I did see the Elvira images. I can't call them pictures as they look like they were done with a six month old 1.5 model. Very fake looking. I have tested the model a bit myself though and was able to get some images that were much closer to photo quality. Just needs better prompts and a desire to actually produce photo quality images. For some reason, Devilmarkus seems to prefer a more Pixar look to a real look. That's fine, to each their own.

ok asshole, did your mom never tell you if you can't say something nice, shut up?

I would call this as a rude, ruthless message with a hint of jealousy, rather than a feedback!

@doodoodaadaa I don't understand what you think the point of feedback would be if all you did was blow smoke up peoples asses and tell them how great they are even when it's not true. I posted a couple of examples of the best images I could get from the model on V1.1. They were decent, but definitely not photo quality, so why would I say they are photo quality. It's fine if photo quality is not what you are looking for. On any version of this model, I have yet to see a photo quality example either posted by the creator or anyone else in the gallery. My sole focus is photorealism, so I have a very high standard which so far I have not seen this model meet. Again, that's ok. Not everyone is going for images that actually look real.

@T8000 Honest feedback is not rude. There are several good models that put out images that look indistinguishable from a real photo at least some of the time. I don't get how I would be jealous of one that so far, as far as I can see, does not. That is just silly and makes no sense at all. 

@AlBat You sure seem bothered by what everyone thinks of your "honest feedback", just an observation. The problem here is really your delivery; this model is not near an "epic fail" in the spectrum of what gets posted on this site. I am personally fairly bothered by some of the garbage I see here on a daily basis, but I don't generally go around saying it to people. It goes to what a previous commenter said about having nothing nice to say.

Add to that your own admission that even the best models only meet your standard "some of the time". AI, and especially SDXL, is not at a level of making consistent results "indistinguishable from a photo." I wonder what you would have to say to the creators of SD, considering the output of the base models. Are they complete morons? Would you have published something so much better if only you could?

Yet another problem with your argument, you have not trained a published model, you have not mixed a published model, and you did not even try this model before criticizing it. This is a classic case of the keyboard warrior troll saying things from behind a screen with nothing to back it up whatsoever. I give automatic credit and generous latitude to the people working on this stuff for no personal gain, and I treat them accordingly.

It is also worth repeating that Markus here has produced some of the absolute best models in 1.5 that I have used (according to my very high standards), and I have tested and used hundreds of photorealistic models in the course of this year. His 1.5 model may be top 5, or at least top 10 of all time in that category. I haven't done as much work in XL, because models like his 1.5 version are still superior to anything I've seen so far in XL, but I have a feeling that this creator will be at the top of those results in the coming months as well.

Here's a little nugget you may not expect from some random on a porno tech site, the actual Secret of Life: Peace is the only thing worth having in this world. If what you're doing isn't leading to peace, you're going in the wrong direction. Find a way to be at peace with what people do in the world, and make what contribution you can towards peace on earth. I promise you will be better off and get more out of this life if you can do that.

@MysteryWrecked The only way we will ever have peace on earth is when all the humans are gone. And to be clear, I don't care in the least what anyone on here thinks of my comments. I am just amused by all the people that get butt hurt over honest and accurate feedback. Again, just saying good things does absolutely no good as feedback. If you already feel, based on people telling you that you are great, that what you have done is great, then there is very little reason to push to improve it. This model is mediocre at best. I got some decent images, but none that look real. Quite a lot of other models I use, do get real looking results. As I have mentioned earlier, not everyone is looking for realism in their results, so for them, this might be the greatest model they have ever seen. If so, congrats to them. Also, if you go back to look at this model at the time I made the original post, you would see that is was much worse than the current version. My initial comment was very accurate at the time it was made. The model has since improved a lot, although still not to the point of realism. Hopefully it will continue to improve. Of course it isn't likely to if people just always say it's great when it isn't.

I'm using this model; the generation is fine up to 98% then at the last the image degrades terribly and the result is terrible. What could it depend on? I hope someone can help me because I really like the model