Styles are trained on synthetic data and do not copy the styles of artists!
Photo 2 - is trained on real data (Dataset from Unsplash)
[M] Merged LoRAs
Recommended LoRA Strength (Weight): 0.8-1
Preview generated without using ADetailer & Hires. fix
Description
Repeats: 20
Epoch: 20
Image Count: 8
Network Rank: 32
Network Alpha: 16
Batch size: 2
Clip skip: 1
Training steps: 1600
Resolution: 1024x1024
Model: ponyDiffusionV6XL_v6
WD14 Captioning
SmilingWolf/wd-v1-4-convnextv2-tagger-v2
LoRA Weight: [0.6-0.8]
This Style is trained on the images created by this LoRA
FAQ
Comments (17)
Is it possible to have a 2d westen cartoon style?
Something like that?
https://civitai.com/models/86546/western-animation-diffusion
@prgfrg23 Yes please
Hi, there,
is it possible to know if you used a particular tutorial to learn how to make style lora?
I do a lot of lora and I'd like to learn how to make styled lora but I'm not sure where to start.
I don't even know how to answer 😐
I just watched other people train models and that's it
@prgfrg23 "I don't even know how to answer"
Bookmarks of videos, or of THE video, that helped you the most. That's what he meant.
@hansolocambo I "literally" don't even know how to answer.
All the videos I've seen I think are almost irrelevant.
I looked up the training information in metadata.
I've done dozens of tests and found the right parameters for me.
I find pictures and caption them with WD14 Captioning
train in kohya_ss and test in AUTOMATIC1111
To be honest, I didn't even think that anyone but me would need these LoRAs.
@prgfrg23
Don't be sorry you can't answer, don't worry. It's just that I'm starting to look into styling loras since the SDXL became available because it's quite complicated to have a consistent style with models like pony diffusion xl but thanks to your loras, it's possible and I thought "hey why not make some too :D".
Quite simply hehe. So I'm doing some tests of my own, so far not very conclusive ;)
I hate to ask what might be a simple question, but what does the "score_9, score_8_up, score_7_up" parts of your prompt mean? I tried searching and couldn't find any info.
I took these tags from the model page
they're responsible for quality
@prgfrg23 Thanks for explaining! :)
score_9 = 9-10\10 quality
score_8_up = 8-9\10 quality
score_9, score_8_up, score_7_up = 7-10\10 quality
Awesome model, do you have social media or discord? I'd love to do paid commission for pony style :)
discord - prgfrg23
Your Style Loras are such a plus for Pony Diffusion.
Thx so much.
Are you planning to create these Style Loras also for the Pony SD 1.5 Model?
https://civitai.com/models/257749?modelVersionId=297064
Details
Files
Line Art Style LoRA XL.safetensors
Mirrors
Line Art Style LoRA XL.safetensors
Line Art Style LoRA XL.safetensors
Line Art Style LoRA XL.safetensors
Line Art Style LoRA XL.safetensors
Line Art Style LoRA XL.safetensors
Line Art Style LoRA XL.safetensors
lineartstyle.safetensors
hf_model_48_07.safetensors
Line Art Style LoRA XL.safetensors
Line Art Style LoRA XL.safetensors
Line Art Style LoRA XL.safetensors
Line Art Style LoRA XL.safetensors
Available On (2 platforms)
Same model published on other platforms. May have additional downloads or version variants.