LoRA trained on 3 Qiyana skins from League of Legends
requested by: @Edenwari
Qiyana: qiyana \(league of legends\), 1girl, short dress, strapless, tiara, necklace, pantyhose, white hair
True Damage Qiyana: true damage qiyana, 1girl, halterneck, pink bra, leggings, jewelry, single hair bun, blue hair, multicolored hair
add "eyewear on head" to add the sunglasses
Shockblade Qiyana: shockblade qiyana, 1girl, headband, jacket, shorts, ninja, fishnets, detached sleeves, fingerless gloves, pink hair, ponytail
It wasn't easy to get everything working in one LoRa, so there might be some blending, but I tried to minimize it as much as possible. I also experimented with multiple models to benchmark the results
models used
anime: YesMix - v2.0, CottonCandy Pastel, Yotta Mix - v4.0, YabaL_Mix V3
2,5D: ReV Animated - v1.2.2, LusterMix - v1.5
realistic: Paragon V1.0, ZavyMix - a2
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 highly 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