MugenMaluMix
Sharing the merge recipe for the Pony Diffusion merge checkpoint MugenMaluMix LoRA mix.
The intended use of this LoRA merge checkpoint is to produce anime and game art character subjects in a fantasy style, and to be used as a base for my future LoRAs.
Quality goals are to prioritise better anatomy (hands, feet particularly) and maintain the base models responsiveness to booru tagging.
📔Resource Overview
🚩Base models:
🖌LoRAs by author motimalu
🖌LoRAs by civitai community:
🧙♂️Merge settings
Merge tool: Supermerger
Settings:
Checkpoint Tuned: AutismMix SDXL
Checkpoint Original: Pony Diffusion V6
LoRA merge ratio, blocks:
akihiko_yoshida_style_pony:0.1,granblue_fantasy_style_pony:0.2,fire_emblem_heroes_pony:0.1,seiken-densetsu-mana-pony:0.1,frieren_nereirfpnxl_pony:0.1,sousou_no_frieren_fern_pony:0.1,Cool_Rance_XL_pony:0.2,Vanillaware_XL_pony:0.1,abmayoPDXL_pony:0.1
🧪Results:
X/Y plot of results for prompts that emphasize each LoRAs style against this merge and its base:
Akihiko Yoshida Style (NieR, Bravely Default, FF Tactics) [Pony]
Granblue Fantasy Style [Pony]
Fire Emblem Style [Pony]
Seiken Densetsu - Mana Series Style [Pony]
Frieren フリーレン [Pony]
Fern フェルン - 葬送のフリーレン [Pony]
Cool Rance XL | Onono Style - Pony Lora by Fish788
Vanillaware XL | Pony Style Lora by Fish788
Abmono PDXL (Abmayo Style) by rred
That's it!
If you want to support the continued development of these checkpoint merges and LoRAs,