It is a problem with Illustrious in general. A non-biased LoRA of the character is the only solution AFAIK. It even does it when the character is not from an anime, the style is included with the character tag and it even bypasses style LoRAs.

@IndolentCat Thank you. But what is AFAIK? :)

@sniffyrt599 As Far As I Know.

@Ikena yes, thats my concern, i could generate stylized semi-realistic faces with v1.2 but now the results' features are too anime style and adding more strength than this messes up faces. so any chance if you upload v1.2 i the future again?

@lunarofleona836 you'll need to use v1.2 locally to generate with it.

TLDR - enjoy getting locked by weird rules like "YOU MUST POST ALL YOUR PROMPT AND SEED" when using NOOBAI variant.

TAKE NOTE THAT THIS RULE APPLY TO EVEN WHEN YOU ARE POSTING ON TWITTER, FACEBOOK, YOU ARE REQUIRED TO POST ALL YOUR PROMPT SEED CFG STEP.

Thats the main reason why this model exist, to create a completely new branch of IllustriousXL.

From https://civitai.com/models/833294/noobai-xl-nai-xl

"III. Open Source Community

To foster a thriving open-source community,users MUST comply with the following requirements

Open source derivative models, merged models, LoRAs, and products based on the above models.

Share work details such as synthesis formulas, prompts, and workflows.

Follow the fair-ai-public-license to ensure derivative works remain open source."

"MUST comply with the following requirements... ...share work details such as synthesis formulas, prompts, and workflows." Users of NoobAI-XL and derivatives are compelled by this license to share "...prompts, workflows". It is a requirement.

noobai is also seemingly released under TWO licenses because there is the license spelled out on the CivitAI description, and then there is the addendum license here: https://civitai.com/models/license/1140829

While the restrictions laid out between the two licenses are similar, they're not the same. It's confusing but the limitations placed are still extreme no matter which way you look at it.

@Madafada1991 EXTREME limitation of POSTING PROMPT? Lol. Read Nvidia limitations.

@Volnovik doesnt matter. plus nvidia limitation doesnt even have anything to do with this. The point of this model is create a whole new branch of Illustrious XL, not related to any NoobAI variant. If you dont like it, leave, Simple. There is a reason why those people that you said "use this model with my lora" actually uses this model and you cant even dispute that.

@Madafada1991 Are you sane? The are 4 versions of my lora: 2 trained on eps and 2 on v-pred. I suggest using eps version of my lora, not use different base model. Moreover, all my images have prompts, settings and resources included (oooh, how could I).

This implies that you made two lengthy posts without even checking content. Shame on you.

@Volnovik dude i do know it. In fact i even skimmed through your own user profile and loras. But thats up to the user what base model they want to use. They can use whatever they want despite your recommendation. Its just a recommendation not forced after all. I even had user uses my lora for pony model which i never intended it to be used that way, do i complaint? No. In fact i was even amused by people's choice. Let people have their own freedom. If you are afraid of incompatibility for v-Pred, put it under SDXL1.0 like the base v-Pred model that L_A_X did. Im pretty sure Epsilon model is suitable for any Illustrious branch, but not v-Pred.

@Madafada1991 aaaand? Eps based lora works better with eps base model? Thats obvious. But people don't want to read. They do not care about eps or v-pred. I saw people use my lora in inefficient way and made a comparison to prove the point. You csn even take a look and decide yourself.

And you bash in with all caps complaining about licensing "issues" of different unrelated base model and making obnoxious claims like me enforcing some base model on someone. Chill.

@Volnovik so there is no point telling them. Let them use whatever they wanted. Hassaku here is providing a model thats completely separate from NoobAI. We dont want any of those "forced to tell workflow" rule. Imagine posting a character on facebook and requiring to post "masterpiece, absurdres, .....(bla bla bla)" while posting. YES , Each picture. Thats how severe that rule is even if they knew they cant implement it due to how many people using AI. So many people have voiced this to him yet he ignored it. This means the rule is intentional. Honestly i dont even want to do CAPS, but the severity of that rule is that high.

@Madafada1991 What is the issue? Just embed it into image. It is automatic for A1111 and variants. It needs one node for comfy. And it highly decreases compute time and prompt fidelity for future reuse of the image in training. That's exactly the reason why it is there and will be everywhere in future.

Anyway, I absolutely admire how you imply that I am enforcing some rules on someone with posting just a comparison by enforcing some odd stuff on me. Thats pure hipocrisy.

@Volnovik unfortunately in facebook it doesnt. As image compression happens. You will need to put the prompt manually, and each of image even. And some other pages like x and instagram too. Compression happens. It is an issue of course. It looks simple. But the hassle of doing it on every image is not worth the effort.

@Madafada1991 As you said previously "doesnt matter". Please stop enforcing your wrong views on people at completely random places.

@Volnovik its not wrong. When you suggest it you are literally telling people to also follows the rule that you dont even know existed before i explained that earlier. Its a domino effect.

@Madafada1991 Oh, by the way, let's check your last image posted:

https://civitai.com/images/53166119

No prompt, no data, no reference to even base model used. This is gating newcomers to the generative AI field. Not just that, not paying respect to people giving results of their efforts to use for free is simply not nice of you. And hey, in the description of the model you are enforcing people to use certain models which you accuse me of doing (which I am not by the way).

Moreover, you are gating stuff under early access but not linking model, which means that base model creator will not get compensation in buzz. That is plain bad attitude.

In other images you have at least prompts, but still do not link models or loaras used to create your promotion images. On civit you can even link them manualy, but you did not do that. That's bad.

And yet here you are moralizing me for whatever reason and enforcing stuff on others.

@Volnovik because its not from any noobai variant. Its from illustriousxl variant. I CAN DO THAT. Thats what separates IllustriousXL with any noobai variant. IllustriousXL came before ANY noobai variant. Go look up the date released. IllustriousXL has different license. Now you didnt read again dont you?

Ofcourse, its up to me to whether use it on early access or not. Illustriousxl license allow me to do that. Unlike noobai license.

@Madafada1991 You doing that is an example of bad attitude and I will point it out. You simply do not respect other peoples work. Period. Moreover you are enforcing your distorted vision on other users and insisting that bad behaviour is acceptable just because you can.

Also you are using caps. Shame on you.

@Volnovik at this point honestly. Its better to tell those noobai dev to fix the license, i have already did that, some others too, you too, should do that.

@Volnovik go on hate me. I have to do this so that people realize how problematic that rule is.

@Madafada1991 you better not tell me what to do based on your limited vision. Especially considering that you proliferate bad attitude in comunity and posting toxic stuff at random places. This is discussion of Hassaku model, not NooBAI. Go there and enforce your stuff there, none of that is needed here. At this point you are just being toxic. And yes, I read the license. And don't see any issue there simply because I do exactly that. And as a novice I had issues with people like you who gate their stuff limiting my capabilities in getting good. And as if it was not shameful by itself you are using caps in the internet.

@Volnovik its not gate. It called expanding your knowledge that to use base model, you are locked under the rule of license. Yes, there is license to use model. And no matter how you bend it, noobai is definitely not for the new you mentioned. Not even you, some sort of veteran knew about the licensing issue, imagine newbie doing it and posting it on social media without knowing that they violated the rule. Which 99% of them will likely do that.

@Madafada1991 Not, you are not only gating new users with ability to recreate images, you are directly undermining their efforts by not linking their models. Models that you use to earn buzz. Every single model has license attached to it. If you wanted to teach others you could make an article with comparisons of different licenses for newcomers. What does apache 2.0 mean? How is it different from Apache 1.0 or MIT? What is wrong with SD3.0 license or BFL license?

But noooooo, you are fighting over minor inconvenience to you personaly, promoting bad attitude among others while claiming that you are trying to do that for community and claiming that you are some knight in shining armor. This is pure hipocrisy.

Especialy coming to a completely different discussions with your claims and using caps.

@Volnovik because its not minor. Its huge problem. Why cant you just ask the dev of noobai to just fix that and we are all good to go. No argument, happy ending.

@Volnovik not linking? Cant you see recommended resources on my models? I put that on every of my model. Thats the model i uses.

@Madafada1991 that's exacly how you deprive creators of that resource from getting buzz and generations tips on this site. By not linking the resource to the posted image. And you are well aware of that if you were allowed to use early access and see yourself as a "veteran" policing stuff. Also you insist on me complaining about something by recommending stuff. Direct quote of you earlier: "But thats up to the user what base model they want to use." Yet you are doing that yourself. Another hypocrisy.

@Volnovik my hypocrisy is not breaking any rules. Thats the differences. You can call me anything you liked, but I dont care about that. All my commissioner paid me properly and its an agree to agree situation. Besides, its not like they cannot use my models, they are just delayed by few days. And the price that i put wasnt even high. Its affordable, and just by farming on this site for a few days you can already get to the number.

@Madafada1991 Your hipocrisy is in claiming that I enforce something by recommending type of my model for people that already use it while doing that yourself.

Your bad attitude is in being toxic and using caps in internet which I think should be a punishable offence.

Your bad attitude is also in doing paid stuff and posting images without linking resources which directly cuts other people's gain in buzz while using their resources for making your loras and images. We both know what and why you are doing.

You are also not posting prompts and workflows on this platform. Did you expect to not notice that when you attack a person who is writing guides about generating stuff? You were given all good stuff to use for free and selfishly try to make gains with it without giving back to community. And you call out other people but don't like being called out for that? Hypocrisy.

You should fix your attitude before moralizing other people on internet, stop whining.

@Volnovik what the fuck are you talking about not giving back to community? Thats like saying the rest of my 200++ lora models that i shared for free is not giving back to community. Im not using those buzz i get for my personal gain. In fact those buzz that i get are used to make even more LORA. Most of my lora have over 2000-3000 steps, sometimes even 7000 and its certainly not cheap to train.

Just because i dont post workflow for one model, and you said i dont post workflow. Go lookup my other model. Imagine only looking up on one model and judging someone from one model while he has made over 250. Plus i am not entitled to post my workflow dude, IllustriousXL license says i dont have to. Im not even using NoobAI variant so I dont need to. No fucking way i can sometimes post workflow when sometimes im on phone posting the Lora. Im even generating it online on other websites.

@Madafada1991 Language, sir. You also use early access feature. Stop using false flags, we both know what you are doing here and why.

Linking model to your images is two clicks in worst case for any image, civitai has that functionality. You can even tag resource that you used if it was other online generation service. I call you out with exact links and one is more than enough. I'm in no mood of browsing stuff of random person that use my comment to spread unrelated propaganda and indirectly undermine my efforts. I literally just clicked on first image. Then scrolled a bit and clicked more. And politely wrote that some of your images have prompts but none have models linked.

@Volnovik fine. I dont have energy to reply to same thing I already said. IllustriousXL is not like NoobAI, and I have ZERO obligation to post ANY of my WORKFLOW. I dont want to post it, and thats up to me. Thank you. You can enjoy your winning trophy and Im out.

j'avais se problème aussi c’était un lora que etait mauvais :/

YES, I WANT TO KNOW THIS TOO

@Ikena Okay! :D Then I suppose V2 has more recent characters aswell?

a man of culture I see