Is this a V-Pred or E-Pred model?

Stellar looking gallery, and not v-pred. Looking forward to trying it out.

Ok now we are talking :) i love creative illust models

Very good model that has both good aesthetic and nice artist styles variety. One problem I noticed is that model struggles with maintaining straight lines in the composition (for example skyline) when it is interrupted by other subject. It happens both with euler and euler_a samplers. You can see the problem right in the 1st image from model's gallery with the girl in the pool. Yoga image has this problem too. Very common problem I remember from sd15 era.

Also can you tell on what illustrious version this model is based? Does noobai tags like "very awa" work?

This model is awesome!

this is one of the best models I've ever used, thanks for sharing. I put it as a reprint in tensor art to be able to use it (my pc doesn't run it locally), and so far, tensor art is having problems, but if you want, I can remove it so you can post it without having to claim it through discord

I was wondering about the recommended resolution—I had heard a while ago that you should stick to multiples of 32px to avoid certain kinds of edge artifacting, but neither 1168 nor 1712 are multiples of 32 (they’re both of 16). I really don’t know the technical reason why there would be such a concern anyway (something to do with VAE encode/decode maybe? I don’t know how this stuff works), but any idea where that wisdom might have come from and why it’s incorrect or doesn’t apply here?

I just found out that your model has been added to Civit online genration, congratulations!

Very nice

wow

Fantastic model, I feel like i am only scratching the surface of what it can do :)

Another great model by Epi. ❤️

Why when I try to do a remix of the image changes the model to illustrious xl and I can't swap it back cause it doesn't appear in the search?

I didn't know where to put this: your models are amazing and I've used some of them nearly every day for almost a year -- no matter what else comes out I keep going back to them for amazing visual fidelity and good ability to follow a prompt/use a lora. And there's nothing to beat opticon 3 for a refiner pass. Thank you so much. For all of it.

An excellent model.
It produces highly compelling images with a strong sense of motion and diverse compositions, while keeping human anatomy consistent with minimal errors.
Its compatibility with LoRAs created for other models (e.g., WAI) is another plus.
It has become my default model.

Extremely great model, can work with many LoRA's at the same time, understands text prompts well and gives really good results without messing up.

The only issue I ran into so far, is that model itself does not contain new characters from games and so on. To make things worse - despite the model working great with many LoRA's at same time, if you add in character LoRA and try to work with it, it sort of works but at same time not really, often making it impossible to make "decent" looking game character or anime character or whatnot if it did not contain data in the model already. So I'd very much love an update with just some chars more added to dataset.

Other than that, 10/10 model, first one I can see myself using for longer period of time compared to any other model tried so far up to date.

What’s going on with 1b compared to the original?

Daa best <3

bro cooked