Update (April 12th, 2023): Apparently this model barely works without a set of embeds. Anyway, here's all of them. As with the checkpoint, it was licensed to me and I'm free to redistribute it. But even if you disagree: different name, different hash == different embeds.
Have fun.
https://files.catbox.moe/0gjduq.zip
--
You may notice a distinct similarity in the output this model generates with a model that was formerly available named 'Illuminati Diffusion'. That is because the weights and configs are identical. However, this is not Illuminati Diffusion v11. That name has been exclusively licensed to one of those shitty SaaS generation services.
In addition, although the weights and configs are identical, the hashes of the files are different. Therefore: different name, different hash, different model.
But still, let's make two things very clear:
Weights and configuration files are not copyrightable, and;
Even if they were, the weights and models were downloaded by me (and thousands of others) under the creativeml-openrail-m license.
That license grants me:
a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable copyright license to reproduce, prepare, publicly display, publicly perform, sublicense, and distribute the Complementary Material, the Model, and Derivatives of the Model.
And no, adding a blurb to the model page that no-one reads does not, in fact, constitute a modification of the license that you uploaded under, or amend the terms of service you were subject to when you uploaded it.
Given the above, I republished a wholly new work containing the original weights. You're now free to use this wildly overtrained model for any purpose you like. I suggest using it as an example of the need for ground truth datasets in training.
Description
FAQ
Comments (88)
So technicly its better to delete the last one and use this one? What does it do if not?
This is just to give it a different hash. The functionality (and output) of Illuminutty v1 and v1.1.1 are identical. You're free to use either one, you'll get the same results.
Ok thanks for answering :P
@ozreac why do you need to change it's hash, tho? just interested. Nvw I've read description. But what is "shitty SaaS generation services."? What is SaaS? They basically stole your model? Or only name?
Oh wait... it isn't yours? It was removed as free model from Civit, and this one like it's clone or something?
Technically he's claiming someone else's model and using a hogwash language. He, ozreac stole someone's model.
@lolmao What exactly did I 'steal'? I have a license for the weights.
@ozreac Correction, you forged the license and you changed the hash. The weights are exactly the same. There's no difference, do you want weights of both models printed on a paper and you calling that a coincidence?
@lolmao .. what did I 'forge'? You can go look at the license on the waybackmachine. And yes of course the weights are the same; that's the point. I only changed the hash because the Civit UI doesn't understand that it's not a duplicate, it's my licensed redistribution.
o7
o7
And it started. Heh.
That's the spirit. May there always be someone in the ML community who isn't hoarding this stuff for the rich. Bravo, no sarcasm.
The Illuminati model disappeared from the mother earth!! Thanks for... republishing? Where did it go? (just curious)
I have to make an update on my review video (94) Mas oscuridad, mas constraste, mejor que Midjourney: Loras, Illuminati y TheAlly. Offset Noise - YouTube
Read description, it's something about shitty SaaS services. I don't sure too.
Básicamente el autor -que ha entrenado el modelo en trabajos que no son suyos (aka no dibujados, no modelados, etc.) osea generaciones de MJ v4 y similares- ha vendido su licencia a terceros (Leo) y ahora no se puede mergear SU modelo porque él lo dice. El chiste es gracioso sí pero es así. Se basa en trabajos de otros y pone límites. La "culpa" la tiene un modelo conocido que puso que mezcló illuminati diffusion pero vamos, si me preguntas a mí. Absurdo. Simplemente eso. Un saludo.
@LDWorksDavid Muchas gracias por la explicación! osea que lo vendió a leonardo y lo sacó del opensourceo jajaja que loco!! Pero y si yo tengo el modelo y lo uso... qué pasa? XD
@stable_diffusion_en_espanol Pues sinceramente dudo que pase nada porque al fin y al cabo estas licencias son eso, un absurdo. La única forma de que cómo artista me dignase a respetarlas sería si han dibujado o entrenado algo suyo dentro y cómo tristemente debo ser de los pocos que hacen eso, pues es lo que hay... Pero vamos, no puedes usar el Illuminati diffusion para crear mixes a partir de él ni tampoco vender generaciones de imágenes que haga ese modelo (ya que cambió la licencia cuándo se vendió a Leo) y aquí entra la pregunta, "sacas algo gratis, cambias la licencia y ya no es tan gratis ni tienes las mismas libertades?" en fin. Ellos sabrán. Como artista y finetunner paso olímpicamente de toda licencia y mis trabajos van a ser "haz lo que quieras con ellos" y ya. Allá cada cuál. Un saludo!
@LDWorksDavid Muchas gracias por la respuesta de nuevo!! Hice un video cortito explicando esto para mi canal! (100) Noticia: Chau Illuminati, hola reemplazo | Stable diffusion en español - YouTube
Y de paso mostré un poco de tu trabajo! Abrazo!
Great stuff. I enjoy it and appreciate your hard work. Not sure about the salty comments.
I'm still not sure, but as I understand, author of this reupload (and it's is just glorified reupload, as i understand), have nothing to do with creating this model and hence with any hard work.
No offense to anyone, tho. I think i understand what happened, but I still doesn't get it how author of reupload get this licensed? As he claims.
embedding https://civitai.com/models/36416/unreals-sd-21-hires
What is the model? Just cats?
@countlippe it is Illuminati merge, that caused author of original Illuminati model to remove if from civit, because model licensed for some site service, and license prohibit using model in merges. Dude refused to remove it, and author of Illuminati remove Illuminati instead. Or something like that.
@SilentPrayerCG Perfect - thanks for clarifying. Messy situation and hard to get my head around.
@countlippe I only started to understand what happened, after googling and asking what happened on NMKD discord server :D
I've added the embeds to the model description.
What happened to Illuminati Diffusion?
It's been exclusively licensed to one of those shitty SaaS generation services.
@zeldasixactual718 sounds like sold-out with extra steps
Wow. Really hates ControlNet:
RuntimeError: mat1 and mat2 shapes cannot be multiplied (77x1024 and 768x320)
Time taken: 0.20s
Torch active/reserved: 3271/3498 MiB, Sys VRAM: 5854/24576 MiB (23.82%)
Does ControlNet work with 2.1?
@countlippe Ah it's 2.1, of course. I should've noticed.
@fiddlesandsticks I was praying you'd found some way of getting ControlNet to work with 2.1!
@HollowBrain I hadn't come across that. Thanks, giving it a shot
ControlNet does work with 2.1 models, I use it all the time. But you'll need ControlNet models for SD 2.1, which are distinct from the SD 1.5 counterparts. Also, in Automatic1111, a setting needs to be changed to use the ControlNet config for SD 2.1 instead of SD 1.5.
Stealing a model and changing hash. Ridiculous
so why this is still online?
CivitAI moderators are sleeping?
.. why would it not be? I’m sharing one of my licensed models with everyone. That’s pretty much the whole point of civit.
Because why shouldn't it be :)
@Gbronski Because commenter is clearly an EXPERT_USER... Can't you read? ;-P
Do you actually 'prefer' that it not be listed?
Is the original creator gonna talk too about the private dataset he used and from where he obtained too?
Because you cannot revoke a license for the exact same work once it goes public. (You can put a different license to a derivative work, but that doesn't affect the license of the original.) That's why. Go home, you need more sleep.
I wish to take no side in this chaos but I am curious if the model copyrighting parties created those models using only licensed material and if they expect copyright enforcement to work out better for them than artists whose art is in those models.
I have a deep respect for intellectual property. No artists are stolen from when a model is trained, and no 'finetuner' is stolen from when I share weights that I have a license to distribute.
Embeddings by chance??
Thank you, friend.
Thanks for posting that back, kind stranger! But just curious, how did you obtain the licence? Illuminati published it first without restrictions and then changed along the way? what happened?
You're welcome. And I downloaded it while it was licensed under the creativeml-openrail-m license, giving me the (irrevocable) right to distribute it. The license never actually changed though, the original maker simply added a blurb to the model description.
I am not a fan of this model, but I appreciate that you are sharing it.
And as the original license allows this, I don´t see any problem. That is the beauty of libre software, that is free even from its creator.
I fully support your effort, I would have no problem with them trying to sell the model... but if you decide doing so after your model was already online for free download you are dumb as hell lol
The link for the embeddings seems to be dead. Could you please upload it again? Thanks in advance!
link is down. can you fix?
It's not down.
So, after a few hours of head scratching on why in god's green earth my inpating was changing the hue as well as the content of the mask, going through dozens of generations, I finally found the culprit. The embeddings. They are awesome for generating, but keep them out of your negative prompt when inpainting.
wasnt the original a 1.5 model too? if so, can you re-upload that one the same way pls? i think i had it as a 1.5 model on my download list when my drive went full and when i had some space again it was gone... :(
Nope there was never 1.5 model
must have saved it in the wrong bookmarks folder then, good to know, thank you :)
The config link doesn't seem to work. Could you re-upload it?
Where should the embeddings go? Neg or pos prompt? one or all??
"N"egative
whats so special about it?
ozreac, I love your model, it is simply magnificent. Congratulations!!! I look forward to your new job... :) (Y)
honor for YOU!! thx for share
crashes when using controlnet, any clue why?
Yes - most controlnet models are for 1.5 models. Illuminati is based on 2.1.
Im also looking for good 1.5 model which is like illuminati.
@talgreen1271 yeah finally realized this was 2.1 and doesn't work with controlnet, bummed out :\
@Davey212 Do you know any other 1.5 model that I can use for controlnet inpaint, instead of illuminati?
mine crash even not using controlnet
Could someone please upload the embeds again? The catbox link does not seem to work anymore.
jesus who cares about copywrights
Yeah, no shit. Training this shit is legally quesitonable in some first world countries.
How do I use this in Automatic 1111? Can someone explain it in simple terms?
sounds like what happened in my pants!
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.

