If you liked the LORA please make a review, it would help me greatly. I have one more LORA uploaded on my profile.

This is style Lora created from images of artist "Yontarou".

95 of 512x768 images are used to train to the point of 16 epoch, on NAI Full Latest model.

With 18 folder loops and batch process value set to 3.

18 x 95 / 3 = 570 steps each steps

570 x 16 = 9120 total steps

Main trigger prompt is Yontarou .

Tags which are recommended to be used to get the body shape correctly:

plump, large areolae, puffy ???????, large breasts

Original image set contains plenty of characters wearing bodysuit, so bodysuit prompt would work pretty well also.

Ghostmix model is recommended to get the best quality imitation (as far as I tested), but should get good imitations also on most other anime style models.