<h2>Check model versions! The Overfitted one may be better for your usecase!</h2><h3> Description:</h3>The funny small lady. This LORA isn't perfect due to the small amount of good data on Maho, it was trained on only 25 images, mostly from the front, and mostly in her signature outfit. This LORA has a bias towards sketchy look, and green trim around clothes. Can be combatted with prompting, but it's not just a plug and play for every type of image. Recommended 0.8 strength. 1.0 is a bit much. Keyword is just "maho" but you can add things like green eyes, long hair, messy hair, etc to further add, and customize the generation. Have fun!

比屋定 真帆 - Maho [Steins;Gate 0]

An over fitted version, more sketchy in style, less prompt variety, but much more accurate maho proportions, use this if you just want Maho and don't want to mess with her outfit or anything like that. The first two images actually use this one!

Files with SHA256: 29f65c6e4cbfe1e87d6f997f9b4ac47250e5f04b3b1ccf4d8c0d21f672a3c43a