LoRA trained on Quinn + Heartseeker Quinn from League of Legends
requested by: @dimpad
Default Quinn: quinn \(league of legends\), 1girl, armor, helmet, cape, breastplate, black hair, short hair
Heartseeker Quinn: heartseeker quinn, 1girl, hair ornament, elbow gloves, dress, ponytail, purple hair, bare shoulders, choker, skirt
models used
anime: YesMix - v2.0, MeinaPastel - V6 ( Pastel ), DarkMeinaSushiMix - v1.0
2,5D: Animesh - Animesh-Pruned V2. ReV Animated - v1.2.2, PastelDiffusedMix - V2.1, MeinaUnreal - V3
realistic: Paragon V1.0
negative embeddings: bad_prompt, negative_hand, ng_deepnegative_v1_75t, bad-artist, EasyNegative
Recommended settings:
Make sure you are aware on the usage instructions of LORA
VAE: vae-ft-mse-840000-ema-pruned (for realistic models & 2,5D), kl-f8-anime2, blessed2.vae (for anime models)
Sampler: DPM++ SDE Karras or DPM++ 2M Karras(euler a for img2img upscaling) from time to time im also using DPM++ 2M alt Karras
Steps: 20 to 35
Clipskip: 2
Upscaler : GitHub - pkuliyi2015/multidiffusion-upscaler-for-automatic1111 4x-UltraSharp(for 2,5D & realistic), 4xAnimesharp(for anime)
CFG Scale : 5 to 9 (mostly 7)
LORA weight for txt2img: 0.5-1 (for the examples i mostly used 0.8 and 1)
Denoise strength upscaling: 0.3 to 0.55
If you want to autocomplete the tags & trigger im using, I recommend using Booru style tag autocompletion for AUTOMATIC1111's Stable Diffusion web UI
For full body shots i recommend using: ComfyUI Impact Pack
or the a1111 alternative GitHub - Bing-su/adetailer
Description
FAQ
Comments (11)
How about ocean song ashe?
I have to check if it has enough pictures. The larger the dataset, the better the chance for you to generate huge girls :D
@AhriMain yeah no problem, you already made most of the popular stuff at this point, maybe you could redo some of the older league loras on this site like anima squad mf or base kaisa which are ancient at this point
@themangamer123 miss fortune xD, I have not found one that feels like the character, in the same way if you do it, it would be great if you did it with a style or 2 at the most since they tend to get confused and mix both sometimes
@dartroe Hmm, I've received this request many times now. I might actually consider it. The two requests that I'm currently working on (Odyssey Kayn and Ortinet Airi) also have a smaller dataset. So it would be good if the character after them had a bigger dataset, where there is more flexibility and higher quality.
@AhriMain I'm really more interested in captain fortune, if you want to do it before of course
You really outdo yourself with each and every LoRA don't you. Incredible as always and works beautifully with your new style LoRA as well.
Ay, thanks! I mean this isn't the best LoRA honestly, Quinn doesn't have the highest-quality pictures, but I did my best to make the most of it. :D
I'll probably attempt to train a checkpoint using a larger dataset out of it. I started with a style LoRA first because it's way faster and to collect feedback to improve the final checkpoint.
@AhriMain Might just be my Quinn bias LOL
any chances for star guardian quinn?
Hi, do you have the training data for classic quinn? I'm trying to make an SDXL / Pony LoRA for her but there's barely any good reference material
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.