Willowlover071 Style LORA, second version. Trained on Autismix, but gives better results in Pony6, than first version, trained on Pony6. Could'nt get more thin details, maybe it's up to sdxl itself, or vae, or something else.

No trigger word, strength vary, to get more smooth picture I lower strength to 0.7

Feel free to experiment and share results