<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]

Has 30% more images compared to the overfitted version, but produces a bit more generic results. Use this version for complex prompts or combining with other LORA's

Files with SHA256: db95b9a75d1e261b8a1b936491896c0c9b1a8c45a0787a290ac555dca6f13bde