Styles are trained on synthetic data and do not copy the styles of artists!
These styles are similar to LoRAs for Pony Diffusion V6 XL, but will differ slightly in names or number of models (Trigger Words may also differ)
Recommended LoRA Strength (Weight): 0.8-1
Preview generated without using ADetailer & Hires. fix
I didn't add Merged LoRAs here because I don't see the point in it. The list of models from which the Merged LoRAs are made
Smooth Anime -- NAI3 Smooth Boys 0,50 + Ani-B - Basic 0,50
Smooth Anime - Night -- NAI3 Smooth Boys - Low Light 0,50 + Ani-B - Night 0,50
Concept Art Ultimatum -- Concept Art Moonlight 0,50 + Concept Art Brush 0,50
Eldritch Oil -- Oil Painting 0,35 + Impressionism 0,35 + Palette Knife 0,30
Gothic Painting -- Gothic Art 0,50 + Oil Painting 0,50
Description
FAQ
Comments (14)
Damn look at you go crankin out these! nice work!
I'm glad you like what I'm doing.
can we get a twilight version plz
Concept Art Twilight This is not a trained LoRA, this is a Merged LoRA, which means that the LoRA is made of several LoRAs
I tried to create Twilight for NoobAI-XL EPS 1.0, but I didn't like the result. for the same reason, there are no Photo styles here. it literally looks terrible!
I would say that the Moonlight model is a replacement for Twilight, yes, the style is a little different, but all these styles for Illustrious are different from the styles for Pony
Some styles look better for Illustrious/Noob than for Pony, but some styles look worse than for Pony. Even though they are trained on the same images
I hope the answer is not too difficult to understand. and I also hope that it is clear why some styles are missing and that there are reasons for this.
I love it when
A) A bunch of styles are consolidated into a single page, and
B) The same prompts get used across the different styles, giving a much stronger sense of the differences, strengths and weaknesses of each one.
Finally, the best pony lora collection ported to IL.
Even if it's not the same quality, i doubt it's worse than the usual IL stuf.
Yes, these styles are different from the styles for Pony. but most of these styles work quite well.
How to use these styles with a different checkpoints? I don't like base illustrious 0.1 checkpoint and with a different ones my images have poor quality and artifacts.
Which models have you used these styles with?
@obuyb404 WAI v12, Hassaku v1.3 A, NTR MIX III, Prefectious XL NSFW, Prefect Illustrious XL. I'm new to illustrious however i use all the recomended settings for each checkpoint and set required strength of your lora with the trigger words (I use ConfyuUI). What i could possibly do wrong? Your Twilight style for pony is perfect, works with different checkpoints and even without the trigger words but with the illustrious your moonlight styles makes images dark, blurry and add artifacts at strength higher than 0.5 (sometimes even lower) at least for me.
@revenantfromshadow I tested LoRA in ComfyUI and didn't notice any problems.
I don't know, maybe there are some nodes in your Workflow that affect quality in this way.
@obuyb404 Are you sure that moonlight works fine with checkpoints i mentioned on suggested strength and also does it work without the trigger words?
@revenantfromshadow
I tested LoRA with these models. WAI-NSFW-illustrious v12.0, NTR MIX XIII, Prefect illustrious XL v1.0
Concept Art Moonlight works well at 0.5-1 strength
I did not notice any artifacts or deterioration in quality. I really don't know what this might be related to in your Workflow.
And I don't quite understand why to use a model without Trigger Words. The style is activated with these words and I trained LoRA with these words.
The list of nodes in my Workflow:
Load Checkpoint
Load LoRA
CLIP Text Encode (Prompt)
VAE Decode
Load VAE
KSampler
Empty Latent Image
Save Image
@obuyb404
Alright, thank you for checking it out.
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.