模型简介: I took a break from the Z fighters and did a beloved side character this time. A couple of trainings, nothing fancy, but it works. I have a kofi if you want to support my work. I also take commissions, so if you are interested on a training, check that out. I also have a Pixiv where I upload pics from testing and tease the things I'm working on! Or go to my wife's Pixiv, she has sexy pics of my LoRAs and WIPs. About the LoRA, his appereance is mrpopo, black eyes, black skin, red lips, pointy ears, no nose, no eyebrows. He's very overfit but is still posible to change his clothes on weight 1 and his gender around weight 0.8-0.9. The Lookout architecture is very imprinted on the LoRA, so just using building as prompt gives similar structures. As always, I'll leave examples in the gallery, I didn't inpaint anything so there's always room for improvement. The source is mostly old, so I recommend using hires with 0.3-0.5 denoise depending on the model. The outfit is: brown naked vest, white turban, blue gemstone, white baggy pants, red pointy shoes, red sash, armlet, earrings, sleeveless.The gemstone color can be changed with some weights. Again, is very overfit sometimes there's no need to prompt all the outfit. Use the part you don't want as negative if is always there. If he gets some kind of top under the vest, use shirt on negative. ~~~~~~~~~~~~~~~~~~~~~~~o(* ̄︶ ̄*)o~~~~~~~~~~~~~~~~~~~~~~~ 示例提示词和参数: Prompt: (8k, best quality, masterpiece:1.2), (finely detailed),solo, detailed illustration,intricate,(upper body), straight-on,facing viewer, a man mrpopo, black eyes, black skin,white turban, brown naked vest, blue gemstone, red sash, white baggy pants, standing at a rocky grassfield, looking at viewer, day, blue sky,arms behind back,smile, no nose,earrings,fat, no brows Negative prompt: bad-hands-5, (worst quality:2), (low quality:2),EasyNegative,lowres, ((nose,monochrome,bad eyes,bad hands,abs)), ((grayscale)) Steps: 30, CFG scale: 20, Sampler: DPM++ 2M Karras, Seed: 3219124994, Size: 512x768, Model: ivory_v13, Version: v1.2.1, CFG mode: Half Cosine Up, Mimic mode: Half Cosine Up, Model hash: 8fc40d9f60, Hires steps: 20, Mimic scale: 7, Hires upscale: 2, AddNet Enabled: True, AddNet Model 1: mrpopo_v0_2(9eaa1d67b12a), Hires upscaler: R-ESRGAN 4x+ Anime6B, Variation seed: 2922820899, AddNet Module 1: LoRA, AddNet Weight A 1: 1, AddNet Weight B 1: 1, CFG scale minimum: 3, Denoising strength: 0.5, Mimic scale minimum: 3, Threshold percentile: 90, Variation seed strength: 0.05, Dynamic thresholding enabled: True, Clip skip: 2 |