Type | |
Stats | 1,448 |
Reviews | (201) |
Published | May 5, 2023 |
Base Model | |
Training | Steps: 7,920 Epochs: 20 |
Trigger Words | nina bodysuit nina bride nina regular nina tekken8 |
Hash | AutoV2 79B6B938A9 |
Lora trained on Nina Williams from Tekken
models used
anime: MeinaUnreal, PrimeMix
2,5D: Lyriel, CamelliaMIx_2.5D_V2, SoapMix, ReV Animated
realistic: fantasticmix_real, henmix_real
Black bodysuit outfit (Tekken 7): nina bodysuit, 1girl, black bodysuit, low ponytail, latex
Bride outfit (Tekken 7): nina bride, 1girl, wedding dress, veil, thigh holster, long hair
Regular outfit (Tekken 5&6): nina regular, 1girl, low ponytail, purple camouflage, blonde hair
Tekken 8 outfit (low sample size): nina tekken8, 1girl, purple dress, leather jacket, sunglasses, short hair
negative embeddings: embedding:bad-hands-5, embedding:ng_deepnegative_v1_75t, embedding:bad-artist, embedding:EasyNegative
Recommended settings:
Make sure you are aware on the usage instructions of LORA
VAE: vae-ft-mse-840000-ema-pruned, kl-f8-anime2, blessed2.vae (for anime models)
Sampler: DPM++ SDE Karras or DPM++ 2M Karras
Steps: 20
Clipskip: 2
Upscaler : img2img in comfyui using 4x-UltraMix_Smooth
CFG Scale : 5 to 9 (mostly 7 and 8)
LORA weight for txt2img: 0.5-1
Denoise strength for img2img: 0.4 to 0.7
If you want the exact same settings, just drag and drop the picture into ComfyUI
for the pictures i used comfyui