!!! UPLOADING/SHARING MY MODELS OUTSIDE CIVITAI IS STRICLY PROHIBITED* !!!
Check my EXCLUSIVE models on Mage.Space: AniMage PXL • AniReal PXL • Lucid Dream • AniMage SD1.5 • Realistic Portrait
SDXL - Pony: AniVerse PXL • AniMerge PXL • AniToon PXL • AniMics PXL • AniVerse XL
SD1.5: AniVerse • AniThing • AniMerge • AniMesh • AniToon • AniMics
Also in Collaboration with Shakker.ai
This model is free for personal use and free for personal merging(*).
For commercial use, please be sure to contact me (Ko-fi) or by email: samuele[dot]bonzio[at]gmail[dot]com
⬇Read the info below to get the high quality images (click on show more)⬇
Aniverse - Flux Dev - make the impossible possible!
The name is a merge of the two words: Animation and Universe (and a word pun: Any+Universe -> Anyverse -> Aniverse)
-> If you are satisfied using my model, press on ❤️ to follow the progress and consider leaving me ⭐⭐⭐⭐⭐ on model review, it's really important to me!
Thank you in advance 🙇
And remember to publish your creations using this model! I’d really love to see what your imagination can do!
Recommended Settings:
With my 2060 and 12gb of vram, with SwarmUI I have an average generation that varies between 100 and 110 seconds
SwarmUI or Forge
Flux Dev: FP8 (work on 12Gb -GPU) or Q8 (but not tested, shoud work on 6Gb -GPU)
VAE: automatic
Width: 768
Height: 1344
CFG Scale: 1
Steps: 20
Sampler: DPM++ 2M
Scheduler: SGM Uniform
EXAMPLE OF GENERAL PROMPT
POSITIVE PROMPT: type of shot, (Subject), the description, (background), more details
EXAMPLE: full body shot, Morrigan Aensland, long green hair, bangs, floating hair, black leather leotard with white fur trim, sleeveless, elegant sitting over a white and soft cloud, (closed legs), bat wings, pink tights with bat prints, black boots, starry sky, detailed, seductive, innocent, (depth of field), (fine textures details)
Do you like my work?
If you want you can help me to buy a new PC for Stable Diffusion!
❤️ You can buy me a (Espresso... I'm italian) coffee or a beer ❤️
This is the list of hardware if you are courius: Amazon Wishlist
I must thank you nuaion and GattaPlayer for their support
You are solely responsible for any legal liability resulting from unethical use of this model
(**) Why did I set such stringent rules? Because I'm tired of seeing sites like Pixai (and many others) that get rich on the backs of the model creators without giving anything in return.
(***) Low Rank Adaptation models (LoRAs) and Checkpoints created by me.
As per Creative ML OpenRAIL-M license section III, derivative content(i.e. LoRA, Checkpoints, mixes and other derivative content) is free to modify license for further distribution. In that case such is provided by licensing on each single model on Civitai.com. All models produced by me are prohibiting hosting, reposting, reuploading or otherwise utilisation of my models on other sites that provide generation service without a my explicit authorization.
(****)According to Italian law (I'm Italian):
The law on copyright (law 22 April 1941, n. 633, and subsequent amendments, most recently that provided for by the legislative decree of 16 October 2017 n.148) provides for the protection of "intellectual works of a creative nature", which belong to literature, music, figurative arts, architecture, theater and cinema, whatever their mode or form of expression.
Subsequent changes, linked to the evolution of new information technologies, have extended the scope of protection to photographic works, computer programs, databases and industrial design creations.
Copyright is acquired automatically when a work is defined as an intellectual creation.
Also valid for the US: https:// ufficiobrevetti.it/copyright/copyright-usa/
All my Stable Diffusion models in Civitai (as per my approval) are covered by copyright.
Description
My personal page for donation (and where you can buy my model in early preview): Samael1976 Ko-fi page
Good morning everyone,
I want to start by saying that this LoRa model for Flux was kindly gifted to me by my friend Franci (gattaplayer).
I’d love to say that it was a collaborative effort, but 99% of the work was done by her, so if you want to donate Buzz, give it to her, not to me.
I only provided the images for the dataset, and she did all the captioning and training.
When she showed me the finished product, I was absolutely blown away... for me, it’s a PERFECT piece of work.
The result is images that are a perfect balance between AniVerse and AniMerge (for simplicity, I'll call it AniVerse).
This gift was given to me to lift my spirits a bit.
You should know that I’ve been considering quitting altogether, or at least taking a long sabbatical from making models and doing training.
There isn’t a specific reason, but rather a series of small factors that have gradually overflowed the cup.
- Lack of new stimuli.
Yes, every model is always different, and you never know how the result will turn out, but eventually, once you’ve learned how to do the training, everything becomes rather "mechanical."
- General disappointment.
Seeing that the AI model user base tends to reward merged models over trained ones has left me a bit disappointed.
It’s especially tough to see people reward 95% clones of AniVerse rather than AniVerse itself, and I find that quite hard to swallow.
Not that there aren’t some amazing merged models, like Photon or Realistic Vision (which I consider absolute masterpieces).
But seeing that cloned merge models (and I emphasize CLONES) that bring nothing new are preferred, after all the work a trainer puts in, leaves a bitter taste in my mouth.
Moreover, many of these clones don’t even give credit to the original models used to create them, which really hurts (which is why I stopped doing it with pony models myself, since others weren’t doing it either).
Oh, maybe I forgot to mention that (speaking for myself here), not everyone may know that creating a trained model takes me about 20 to 25 days of work (between building the dataset, captioning, tweaking, and training).
While for a merged model, if you’re lucky, it can be ready in a couple of hours (including all the tests needed to ensure it works well).
If things go badly, really badly, and I can’t manage to find a good initial recipe, a merged model has never taken me more than 3 to 4 days.
- Lack of respect.
Seeing my models uploaded every single time on sites outside of Civitai is, for me, a sign of disrespect for all the work that goes into creating these models.
Every single day, I waste time claiming my models (to have them removed) from unauthorized sites like Pixai, which profit off the work of model creators to make money (and they make loads, believe me) without giving any recognition to the creators themselves.
Just to give you an example, my last electric bill (which in Italy can be monthly or bimonthly; mine is monthly) for July was €235.
True, it was partly due to the air conditioner, but you can understand that having a PC running 24 hours a day for 25 days at home doesn't help.
On the other hand, I want to sincerely thank, with all my heart, everyone who has shown me support and appreciation for what I've done so far, even if it was just taking 3 seconds to write "thank" or even just an emoji.
Believe me, it means the world to me, more than anything! It’s you who have kept me going with this "work," holding on and pushing forward. If it weren’t for you, I probably would have quit long ago.
So, really, THANK YOU!
This long message isn’t meant to start any unnecessary drama; that’s not what I’m about, and it’s not the goal.
It’s to help you understand the reasons that have made me reflect on my future path.
So, if at any moment I stop publishing models, it’s probably because I’ve chosen to take a break from all of this and cool down a bit.
But even though I’m considering taking a break, I don’t rule out coming back more motivated than ever.
That being said, know that I’d like to bring AniToon and AniMics as LoRa for Flux as well (I’m unsure about AniMerge), but I’ll need Franci's (gattaplayer's) help with that.
See you soon!
Buongiorno a tutte/i
Voglio dire fin da subito che questo modello LoRa per Flux mi è stato gentilmente regalato dalla mia amica Franci (gattaplayer)
Mi piacerebbe dire che sia frutto di una collaborazione, ma il 99% del lavoro lo ha fatto lei, perciò se volete fare donazioni di Buzz, dateli a lei non a me.
Io le ho solo fornite le immagini del dataset dopodiché lei ha fatto tutto il lavoro di captions e training.
Quando mi ha mostrato il lavoro finito sono rimasto a dir poco sbalordito… per me è assolutamente un lavoro PERFETTO.
Il risultato sono immagini che sono una perfetta via di mezzo tra AniVerse e AniMerge (per semplicità lo metto come AniVerse).
Questo regalo mi è stato fatto per risollevarmi un po'.
Dovete sapere ho preso in considerazione di mollare tutto, o almeno prendermi una lunga pausa sabbatica da fare modelli e training.
Non c'è un motivo specifico, ma tanti piccole gocce che sono andate a far strabordare il vaso.
- La mancanza di nuovi stimoli.
Si, ogni modello è sempre differente e non sai mai come sarà il risultato, ma alla fine, una volta imparato a fare il training diventa tutto piuttosto "meccanico"
- Delusione generica.
Vedere che la base degli utilizzatori di modelli AI va a premiare modelli merged piuttosto che trained mi ha un po' deluso.
Soprattutto vedere persone che vanno a premiare cloni al 95% di AniVerse piuttosto che AniVerse stesso, per me è abbastanza difficile da digerire.
Non che non esistano modelli merge stupefacenti, come per esempio Photon o Realistic Vision (che per me sono assolutamente dei masterpiece).
Ma vedere che si preferisce modelli merge, (sottolineo CLONI) che non portano nulla di nuovo, dopo tutto il lavoro che un trainer fa, lascia l'amaro in bocca.
Oltre il fatto che tanti di questi cloni, non riportano neppure ringraziamenti ai modelli utilizzati per crearli, fa davvero davvero male (motivo per cui con i modelli pony ho smesso di farlo io stesso, visto che gli altri non lo fanno).
Ah scusate, forse mi son dimenticato di dire che (e qui parlo per me), che forse non tutti sanno che per fare un modello trained mi vanno via dai 20/25 giorni di lavoro (tra creare il dataset, captioning, risistemare e training).
Mentre per un modello merge, se si ha fortuna in un paio d'ore (compresi tutti i test necessari per capire se funziona bene), è pronto.
Se va male, ma tanto tanto male, nel senso che per sfortuna non si riesce a trovare una buona ricetta iniziale, un modello merge mi ha portato via non più di 3/4 giorni.
- La mancanza di rispetto
Vedere che i miei modelli ogni santa volta vengono uplodati in siti al di fuori di civitai per me è una mancanza di rispetto verso tutto il lavoro che c'è per creare tali modelli.
Ogni santo giorno perdo tempo ad andare a reclamare i miei modelli (per levarli) da siti non autorizzati come Pixai, che approfittano del lavoro degli autori di modelli per fare soldi (e ne fanno a palate, credetemi) senza riconoscere nulla agli autori stessi.
Tanto per dire, la mia ultima bolletta elettrica (che in Italia può essere mensile o bimestrale, io ce l'ho mensile) di luglio è stata di 235€.
Vero che è stato a causa anche del climatizzatore, ma capite che avere un PC a casa che riscalda e che funziona per 24h al giorno per 25 giorni, non aiuta di certo.
Voglio invece ringraziare di cuore, con il mio massimo affetto, tutti coloro che mi hanno espresso vicinanza e apprezzamento per quanto fatto finora, anche spendendo 3 secondi scrivendomi "thank" o solo un emoticon.
Credetemi, per me conta tantissimo, più di tutto! Siete voi che fino ad adesso mi avete fatto stringere i denti e continuare con questo "lavoro", probabilmente se non fosse per voi, avrei già mollato da un pezzo.
Perciò, veramente, GRAZIE!
Tutto questo mio lunghissimo pippone non è per creare un flame inutile, non ci tengo e non è questo lo scopo.
Ma per farvi capire le motivazioni che mi stanno facendo riflettere sulla mia strada futura.
Per cui, se da un momento all'altro non pubblicherò più modelli avrò probabilmente scelto di prendermi una pausa da tutto questo e sbollire un po'.
Ma anche se sto considerando di prendermi una pausa, non escludo di tornare più motivato che mai.
Detto questo, sappiate che mi piacerebbe portare anche AniToon e AniMics (su AniMerge sono in dubbio) come LoRa per Flux, ma qui avrò bisogno della consulenza di Franci (gattaplayer).
A presto!
FAQ
Comments (25)
You are so motivated, Samael! Hahahaha I still haven't tried FLux (I'm too lazy to install everything, free up space, etc, hahahaha), and I don't know if it will run on my PC.
I just saw that if you want test, you can use civitai generations, without install anything ;)
Always grateful for the incredible work you do for us!
Thank you for you words, but this time all the merits go to gattaplayer 🤗
You always do great stuff from SD1.5 to this ... Thanks
Thank you so much, but like I said, all the merits go to my friend gattaplayer
Thanks for sharing a lora first.
With Flux model sizes, it's just not feasible to download all the checkpoints we like.
And also everybody now has a specific Flux container / format they prefer: UNET, NF4, GGUF ( Q4, Q6, Q8 ) ...
With a lora we can get the look we want, using less disk space, and a model format that's better for our system resources.
I think loras will gain popularity over large base models, in terms of downloads.
Seems to be working with Schnell!
good to know! thank you! I've not tested on Schnell! Thank you!!!! ❤️
@Samael1976 a proposito, hai visto che ho fatto un lora della Panda per AniVerse 4? XD
Thank you for the hard work you put into these! I've always enjoyed your models ^_^
Thank you so much! This is the kind of comments that give me the strenght to not give up 🤗
As always, TNX!
Thank you my friend
Given how much heat a GPU creates while working with generative AI, at least some of that air conditioning can also be considered a financial cost for model training. I'm not kidding.
Totally true!
Will there be ones for AuraFlow (PonyFlow) once they break out a v1? Can't just let Flux have everything y'know
Honestly, it's not planned at the moment. Simply because I can't do local training of flux locally with my current PC. (Theoretically, I should get a new PC at the beginning of 2025 with the release of the 5090 rtx). So spending money to do a training on RunPod (again) without having the certainty that the training will be successful, (last time I threw away €70 with the result that I didn't get the results I hoped for, throwing everything away), is not an option.
I love this lora. Thanks for the great work !
Thank you my friend!
Curious about the sorts of characters available on this model. Is it similar to your other models? And would there be a list for them?
This is quite a long discussion. Flux is very resource-intensive and the training was done on a limited sample of images from the dataset (500). So it's more no than yes.
Dearest civilians! It's finally coming... if you're curious, in my profile there's "hidden" (but not too much) something new. If you find the PDF, feel free to share! and of course if you don't want to miss the news when it comes out, I suggest you follow my profile on civitai.
Guys, OmniA is out! The model I've been working on for 6 months!
Please go download it, it's very important to me
Let's save a great artist!
I met a super talented person on Instagram who was ready to stop creating due to a lack of support. I talked her into trying for one more month, and this is where you come in!
Could you please take a moment to follow her? Let's show her that her awesome work is appreciated!
Instagram: https://www.instagram.com/luna09152004/?g=5
Patreon: https://www.patreon.com/c/Luna09152004/posts
A huge thank you to everyone who decides to follow her journey!
Details
Files
Available On (2 platforms)
Same model published on other platforms. May have additional downloads or version variants.










