Nova Pixels XL
Nova Pixels XL is SDXL model that is focused on pixelated / sprite arts
All images on example were created with diffusers with custom png tags
Rules
You cannot use the generated images for commercial use if it's not edited (or just turning it to black and white)
You can share images without any restriction if you don't monetize it
Advertising this model to outside is always welcome
Recommend Settings
Sampler: Euler a
Steps: 20~30
Clip Skip: 1-2
Denoising Strength: 0.65 - 0.8
CFG Scale: 4~6
Prompt: masterpiece, best quality, amazing quality, 4k, very aesthetic, ultra-detailed, (pixel art, dithering, pixelated, sprite art, 8-bit:1.2), {Prompt}, BREAK, depth of field, volumetric lighting
Negative Prompts: modern, recent, old, oldest, anime, illustration, cartoon, graphic, text, painting, crayon, graphite, abstract, glitch, deformed, mutated, ugly, disfigured, long body, lowres, bad anatomy, bad hands, missing fingers, extra digits, fewer digits, cropped, very displeasing, (worst quality, bad quality:1.2), sketch, jpeg artifacts, signature, watermark, username, signature, simple background, conjoined, bad ai-generated
Description
Initial Release
NoobAI EPS v1.1 + Illustrious v2.0 stable DARE applied
FAQ
Comments (13)
Ok, now you've got my attention
Pixel Art style model by my Favorite model maker.
Dude. Thank. You. <3
This could be super useful. I hope it can do other things than just anime girls :P (I know I'm in the minority here, but I don't care for them)
I can confirm, it does furries really well.
@NeonN3K0 I second that in motion. Gotten some good images so far with it.
Good job, but I still don't get it. Why would I download a model that weighs 6 GB for the sake of pixel style when I can just use some Lora that can make pixel style and can weighs about 200 mb? Or just use prompt "pixel style"?
You can use a lora without a checkpoint?
Lumira17 I can use any model that combines many styles and use lora for a pixel style.
djdiii Good luck out there lol
To tackle this question more honestly,
A pixel style lora is using a (typically) much lower dataset involved, and it's quality is going to vary a lot depending on what Checkpoint you're using it on. Some checkpoints do cover a wide variety, that's true, and there's nothing stopping you from doing it that way. But just like the reason you'd use a 3D, realistic, or 2.5d checkpoint to get much better quality and variety of those types of work, using a pixel style checkpoint which has a LOT more going into it, is going to give you a much larger scope of results.
And beside that, those 'pixel style' loras that weigh about 200mb? The best thing is that you definitely could still be using those to tune the pixel style of a pixel-art focused checkpoint, in the same way as you'd use realistic style loras on those realism checkpoints.
That more or less explains the 'why', and it's all up to your preference in how you generate your stuff. Give it a shot, you might be surprised, or if you have a good 'cover all' checkpoint you wanna stick with, then you can save 6gb and it's no sweat.
Hope that helps!
Love it! Would it be possible for you to make a lower-pixel version, or a version where we can control the pixel size?
This model works extremely well! Love it!