SDXL LoRA trained with Civit-AI trainer for Pony V6/Pony merges, trained on the style of jotun22. They are a furry artist who is known for their feral work with anatomically correct bodiesJotun22’s subjects are generally feral creatures, predominantly equines and other hooved types. Though they do the occasional dragon and centaur. They have some crossover in MLP fandom, though their art is more realistic than cartoony. For the LoRA, I’ve found depending on the goal it might be best to keep source_pony out for more realistic horses and source_pony in for more cartoon proportions. I’ve included examples of both in gallery. While Jotun’s art does apply more to feral body shapes it’s still possible to prompt for anthro figures in his style.Starting with Eular A seems the better sampler, though YMMV.For trigger words, jotun22 or by jotun22 will work, and adding source_pony optional depending on goal.A start of a string might look like: &lt;lora:Jotun22_Style_PonyXL:0.7&gt;, jotun22, by jotun22, score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up, source_furry, source_pony, rating_explicit,

Jotun22 Style [PonyXL]

Expanded data set. Increased step training

Files with SHA256: 40f3d613bd1e62bf0b3dcb29642ec18f5a3faaed27d8e23478d7346f98fda938