This model has been uploaded with the expressed permission of the author, epiTune. I do not take credit for the creation of this model.
A while back, epiTune and I had been talking about models that I had liked - ones which seamlessly blended some of the best aspects of cartoon and anime with photorealism. However, 2.5D didn't quite scratch the itch that I wanted, and I often used base anime models, and refined with photorealistic ones. I always adored the poses, angles, and fantasy aspects of cartoon and anime models, but they lacked the intricate detail and unique faces that photorealistic ones offered. Out of nowhere epiTune asked me to provide feedback on dozens of photos for a checkpoint he was working on, then sent me the first version of this model soon after. After several iterations, and further feedback, Pzeudo was the result - it was never released to the public, until now.
He expressed to me a couple of days ago (as of writing this) that if I wanted to release the model, I may. After initial hesitation, and some thinking, I feel it is appropriate to do so, with the understanding that all credit truly should go to epiTune. My effort in this model's creation are so minimal, I wouldn't feel comfortable taking credit away from them.
So now, I present to you epiTune's Pzeudo.
What is Pzeudo? Some may label this photorealistic, but it's not. Some may label it as 2.5D, but, again it's not. It's neither. It's like 2.75D+, or fantasy photorealism - it's that perfect blend between, like something you'd see in a doctored fantasy photograph, or an abstract artistic movie. Pzeudo is insanely creative, but grounded in visceral realistic fantasy, with vivid colorful scenery, otherworldly spaces, and the grace of a cinematographer. It's the perfect border realm between the mind's eye and reality, with the artist only being held back by their imagination. Whether making moody crypts for denizens of the dead, or crafting a majestic warrior looking upon a valley on a starry night, Pzeudo bursts at the seams in a lived-in world where you're the master.
What is Pzeudo great at?
- Effortlessly creating vivid, majestic, and lush scenery
- Cinematic, moody, and lived in worlds
- Beautiful characters with lifelike faces
- Tangible and terrible monsters, creatures, and otherworldly beings
- Stunning atmospheric lighting, with vivid and balanced darks
- Fantastical but grounded abstract imagery
- Very high detail with little effort
- Easily does NSFW
These are the recommended settings, based upon epiTune's models:
- DPM++ SDE Karras
- CFG 5.0 - 30 Steps
Recommended Prompts:
Positive Prompt
score_9, score_8_up, score_7_up, ...Negative Prompt
score_6, score_5, score_4
Additional Optional Prompts:
(This is based upon experience using this model daily, and are completely optional, and I think bring out some of the best qualities of the checkpoint. They are not in order, not necessary, and can be mixed, matched, combined, or otherwise):
Positive Prompt (Optional)
masterpiece, HD, great angle, fantasy, abstract, filmic, cinematic, atmospheric, mystical, magical, detailed background, realistic, stunning, amazing angleNegative Prompt (Optional)
low quality, bad quality, low detail, cartoon, animeDescription
Initial Release (likely to be the only version)
FAQ
Comments (15)
It looks so gorgeous! 😍
Thank you! EpiTune knocked this out of the park. I've been very fortunate to use it, and further, give the public access.
@TurinBjorn Indeed, thanks a lot for releasing it! (BTW: don't forget to switch off the "sell generated images" and "sell this model and merge" if you don't allow it, and the same with "share merge of this model")
@n_Arno thanks! I went ahead and did as you said.
@TurinBjorn Perfect! People can still ignore those, but it is better to clearly state what you authorize or not, this way, they can't say they weren't aware 💪
@n_Arno Disgusting, when mergers restrict their models. And they not even mention, what models they merged.And they use full unrestricted finetunes. You can fuck off.
@anikansel290 Indeed. Most of the time, the source models if they allows merges ask for the same permissions => so, allowing merge too. Merging stuff and then disallowing merge is not fair play. But i can understand if some more work than just mergse has been done: direct finetune of the resulting merge or a merge of its own trained LoRA does count as "original" work on the checkpoint.
Anyway, what is clear: providing source link to the original models is the bare minium.
I don't know how to describe it, but I think this model has an extraordinary light and shadow performance.
That’s one of the things that struck me about it as well. Previously, I had to use a bunch of models and LoRAs to achieve this, but epiTune somehow balanced that vividness and a great spectrum of light that just hit so sweetly. The depth is just so satisfying and pleasing. Thank you for mentioning this!
It looks amazing on all aspects but i can't add this ressource anywhere on civitai (not found). Any clarification would be appreciated, thanks.
Add this resource? You mean, like generate with it on the website itself?
Is this the model that you did that article about the effects of different words in prompts? If so, I would enjoy a lot of those image (and there were a lot) to be added somewhere to be seen and enjoyed
Most of the images made for the article were resized, with the originals deleted, since there were hundreds. Sorry! However, I’ve made many images for the model, shown here.