[v3.1 is still in further testing. Updates regarding new findings will be updated in the "About this version" section]
UrangDiffusion v3.1 (oo-raw-ng Diffusion) is the first UrangDiffusion version that utilize Animagine XL 4.0 as the base.
The name “Urang” comes from Sundanese, meaning “We/Our/I.” The history behind the name is to make the model not only suitable for me but also for many people. Another reason is that I use many resources (training scripts, dataset collecting scripts, etc.) from other people. It’s unfair to claim this model as “my sole work.”
Standard Prompting Guidelines
Prompting guide:
Default negative prompt:
lowres, bad anatomy, bad hands, text, error, missing finger, extra digits, fewer digits, cropped, worst quality, low quality, low score, bad score, average score, signature, watermark, username, blurryDefault configuration: Euler a with around 25-30 steps, CFG 5-7, and ENSD set to 31337. Sweet spot is around 28 steps and CFG 6.
Training Configurations
Finetuned from: Animagine XL 4.0 Base (NOT 4.0-Zero)
Finetuning:
Dataset size: ~1,600 images
GPU: 1xA100 80GB
Optimizer: AdaFactor
Unet Learning Rate: 1.25e-6
Text Encoder Learning Rate: N/A (Turned off)
Batch Size: 48
Gradient Accumulation: 1
Warmup steps: 5%
Min SNR: 5
Epoch: 15
FAQ
Q: Images are sometimes noisy.
A: This is a common issue with Animagine XL 4.0 models in general. The base model is trained with only 10 epochs, which lead to the model being undertrained. Unlike Initial N or Initial I model that are trained with more resources.
Q: Hires fix model?
A: Check out the cover image metadata, you'll find it there.
Q: Initial N/Initial I is better.
A: Just leave and do not use the model. Simple. No need to announce your departure. Except you're willing to leave a constructive feedback or willing to fund future projects.
Special Thanks
My co-workers(?) at CagliostroLab for the insights and feedback.
Nur Hikari and Vanilla Latte for quality control.
Linaqruf, my tutor and role model in AI-generated images, and also the person behind tag ordering.
License
UrangDiffusion v1.0-v2.5 falls under the Fair AI Public License 1.0-SD license, while v3.x falls under the CreativeML OpenRAIL++-M license.
Description
Using huber loss for better training.
Using random cropping for better image results.
Added new characters from new series, updated undertrained characters, and add new characters from existing series.
FAQ
Comments (6)
Im curious about how you say "masterpiece" quality tag is interfering with finger generation and so you uses "best quality" tag instead,
lets say if I still put that "masterpiece" tag while still having "best quality" tag, does it affect anything?
It'll broke some detail still. You can take a look at this post: https://civitai.com/posts/8469534 and maybe try it yourself.
That tag is problematic for a number of SD1.5 anime finetunes as well. I think everyone training finetunes are doing something wrong.
Hello, I like your model, thank you for the update to 2.0 My question is. What would be the reccomended parameters like Denoise Strength, HiRes steps and the upscaler for HiResFix? thanks in advance.
Hi, my friend Raelina is a more expert than me in this, so I'll just quote them:
Hires.fix Setting:
Upscaler : 4x NMKD YandereNeo XL
Hires step : 10-15
Denoise : 0.1-0.3
===
Hope this helps!
@kayfahaarukku Thank you so much for your answer., I do appreciate it. I'll be sure to try it as well.