The new version is for Klein9B. thanks to Klein's ultra-fast response speed, image editing has entered a new era. The training of this LORA was full of twists and turns, but I was lucky to get this version, which has completely surpassed my previous Qwen version. I call the new version "my nano banana moment". Friends who have experienced its performance in advance should understand what I'm talking about ( for those who are not sure, you can click here to learn more first ). I won't say more. I wish you all a great time!
Experience the A2R Sponsored Version here ( online workflow ) 👈
want contact me👇
X : @AIAVJLeo | [email protected]
| 🙈🙉🙊OK, that's all
If you like my work, please like, follow, and share. This will help more people! Thank you.
English is not my native language. The above content was translated by artificial intelligence. Please forgive me if there are any inaccuracies.
Finally, have fun! AI never sleeps! Peace!
Description
FAQ
Comments (36)
Wow this is easily the best LoRA for this type of thing so far. Really good on Klein 9B.
Thank you for your recognition. If possible, please post some images so that everyone can also feel your "Nano-banana moment".
is the klein example here using the standard or the sponsored version? 😅.
I’m guessing standard, unless we’re seeing “standard (sponsored)” example now 😆
Both are included. I conducted hundreds of sets of tests and only released it after confirming that it is a stable version. However... I can't remember exactly which ones are the standard version and which are the sponsored version. Sorry.
Very good klein comparison examples! I hope it's not a gated lora though.
to be honest, I wouldn't say it's an ungated LORA
@vjleoliu well you do you. But I think releasing undertrained loras for free and the actual tested lora gated is trash behavior for an open community member. Not saying you did that, i'm not that interested in doing this research.
@diogod Oh, that's what you meant. Sorry, I misunderstood you. I thought you were asking if this model contains pornographic content. LOL, by the way, the model I released is actually a fully trained one. I don't select the best version from a single complete training session to set as paid and the second-best as free. On the contrary, I first conduct multiple training sessions, then pick the best version to share in the community. Based on this version, I increase the training parameters, such as rank (if you want to know), and retrain a new version. If you have also trained models, you should know that this means requiring more computing power and more time. Therefore, I also need someone to fund me to train better versions. This is roughly my approach, and I hope it won't cause new misunderstandings.
@diogod the ones he releases for free are fantastic, they are not undertrained,
@vjleoliu good to know, they do look great. Thanks!
@1962921824 Thank you for your recognition. If possible, please help like, follow, and share so that more people can play with it.
@diogod Thank you for your recognition. This shows that my efforts have been somewhat effective. If you like it, you can give a free like, follow, and share to let more people feel its fun.
searching the workflow on patreon... "Now, join my Patreon Tier 2 membership, and you can get it for free".... bro, if i must paid, it's not free....
Oh, yes, you're right. Actually, what I meant to say is that once you become a member, you can get many resource packs that require payment for free, not just A2R. But obviously, there was a problem with my wording. I've been using artificial intelligence to assist with translation, and this can indeed lead to situations where the words don't convey the intended meaning. Thank you for your reminder.
@vjleoliu You should use something like "with no additional cost"
@Shurik Oh! Right! That sounds more like what I wanted to express. Thank you for correcting me!
looks great, I love that it seems not to change the eye-size and proportions too much.
I always want to change drawings to real, but keeping the exact drawing details, but edit models typically try too hard for the realism and change composition.
If you were to update this, it would be great if you could make it even more-so not-change the proportions.
Yes, as you can see and expect, this is how I conceived it at the beginning of the training, so it can better restore the exaggerated body movements and perspective composition of anime.
@vjleoliu thanks, I absolutely love that. This is the best such lora I've seen.
If you do ever make a V2, I'd love even more in that direction (like, this only does the tiniest amount of not-exaggerating, if that got closer to 0 it'd be even better).
@ForeverNecessary737716 thx, bro, If you like my work, please like, save, and share it. That way, the system will automatically notify you when there are updates. Alternatively, you can join my Patreon channel to get first-hand resources.
Not sure what I'm missing but it's just making 2d video game characters asian.
请问klein版的工作流可否提供一下？我似乎跑不出例图的效果
图像给我，我帮你试一下
Hey, is your patreon down? I keep trying to view your content be it just says page has been removed
Unfortunately, yes, my Patreon has been attacked. You can follow my X, and I will continue to update there.
Hello! What workflow do you use? If you use ComfyUI. Thanks for your effort!
I updated the introduction
@vjleoliu I don;t see any mention of shared workflows in the introduction, or in your linked article. Is this all locked behind a paywall along with a "better" version of the Lora as well? And if so what are the differences between Lora versions? It would be much more encouraging to pay for something if there were clear answers to these questions.
@Boobooberry The workflow has been released on the Floyo platform, as stated in the description. Ever since my Patreon was hacked, the paid version has no longer been available. Except for the Floyo platform, the LoRA version there is the only Patreon-exclusive version across the entire internet.
I created a workflow that compared the lora effects of multiple anime-to-real human conversions, including this lora（A2R_Klein_Standard）.You can click the link to give it a try.
Thanks to the author for providing such a powerful tool for our community!
Flux Klein is just from out of this world...crazy model.
If you use a negative strength, will it turn real photos to anime? 🤔
Klein version named «Standard». Is there a non-standard?
How many people want the sponsored version? Please leave a comment here.
I used to have it and lost the file, I got it when I paid for your patreon, I would love to have it back
+1