Flat Colour Anime!
A vibrant anime style trained using a high quality dataset captioned in natural language. This is my labour of love and I'm constantly working on improving it.
Please consider following to be notified on updates to this style and many more interesting styles I create ❤️
Important notes below! Please read
Keywords:
Check the version details on the right for the correct trigger words for your version. Currently these are:
"Flat colour anime style image showing" for Flux Dev and Schnell, and Z-Image Turbo
"fca_style" for Pony
"fca style" for SDXL
If you don't use the keyword, you won't get the most out of the style.
Version Key:
✅ Current
⛔️ Deprecated
Select the relevant version for your model from the top.
Update notes for version 3.4:
Now available for both Flux Dev and Schnell
Longer training time with a lower learning rate
Trained at 1024px vs 512px
Improved dataset quality
Added regularisation images to the dataset
Description
FAQ
Comments (17)
This amazing lora on top of Neta Art is perfection, to me. The flux version is looking good too, so far. Thank you!
Thank you!
@CrasH I guess I do have one complaint/question. Do you have any advice for using the trigger word without causing my characters to especially enjoy wearing "FCA" t-shirts?
@bc732 Yeah, if you use version 3.4 then the trigger words are changed to "Flat colour anime style image showing..." which stops it from showing "FCA style" everywhere 🙂
How did you go about training Flux Schnell?
I use AI-Toolkit. There's a flag you set in the config which switches it to train Schnell instead of Dev.
@CrasH Oh okay O:
I'll have to look into it, I use Schnell lots but there aren't a lot of loras for it, and Civitai isn't set up to make them either. How fast does training go with it? Is it faster training or similar to Dev?
@LemonSparkle Same speed. 3000 steps on a 3090 @ 512x512 is about 3 hours.
@CrasH I'm still kind of a noob, I had everything setup for SD 1.5 like a year ago, did a bit with it, did a little local and some on colab, all command line, but kinda lost track of everything and forgot basically all of it, especially since pretty soon after that I could gen stuff for free online easier than anything I could do local or even colab.
I mean, my home hardware sucks and would probably take at least 10 times as long as that for me, but I have a spare computer just sitting attached to my TV doing nothing like 99% of the time so it wouldn't really matter if it even took a few days or even a week I guess, if it meant I could get some more Schnell stuffs to play with when it was done, cause I like Schnell, especially when letting it do a little more than 4 steps.
I'm newer to flux, do they go in LORA's folder or checkpoint?
Some of the images I generated contains signature, is there way to fix that with prompting?
Not that I know of, sorry. Is it just this Lora you're using or are you combining with others? As far as I know this one shouldn't introduce extra signature bias.
"Flat color anime," and more than half the images in the gallery are 3D or airbrush style. Some people... 😅
To be fair, thats nothing weird if you mixing loras. I used loras that may not really doing what the loras supposed to do, only added to change the style ever so slightly. And so happens that when you upload, it adds in that gallery as well.
This is one of my favorite Pony loras. Any chance for an Illustrious/NoobAI version?
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.