Development of this model has been completed.
Please use this one from now on.
こちらのモデルは開発終了しました。
今後はこちらのものを使用してください。
FiaMix Reboot Series
FiaMix Reboot
https://civarchive.com/models/161341?modelVersionId=211986
FiaMix Reboot H
https://civarchive.com/models/176353/fiamix-reboot-h-nsfw?modelVersionId=213535
Description
More erotic elements have been added.
FAQ
Comments (3)
Excuse me, I have a stupid question, I'm a little new. Is each version independent? I mean, if I download an older version of the model, will I be able to get the same results, or will they be completely different because they are different versions?
All conditions must match to produce exactly the same thing.
Here is the information on what I used.
I think this will produce something similar.
--
parameters
(pornography), (anime,8k,masterpiece, top quality, best quality,beautiful and aesthetic:1.2,professional illustrasion:1.1,ultra detail:1.3,perfect lighting), extremely detailed, highest detailed, incredibly absurdres, highres, ultra detailed, intricate:1.6, (girl), noble, (highly detailed beautiful face and eyes), pale skin, parfect body, thighs, large breasts, big areola, pink eyes, elf ears, pink hair, long hair, straight hair, long skrt, thin pubic hair, oily skin, cute,
bare shoulders, china dress, chinese clothes, short dress, cleavage cutout, sideboob, thighs,
Negative prompt: (worst quality, low quality:1.4,nomal quality), watermark, logo, missing limb, username, watermark, signature, bad anatomy, buttock hair, EasyNegativeV2, negative_hand-neg, (((futanari:1.4))), (red cross:1.4),
Steps: 40, Sampler: DPM++ 3M SDE Karras, CFG scale: 5.5, Seed: 3502419437, Size: 680x960, Model hash: 9b8888bbe0, Model: FiaMix++H V5.0, VAE hash: c6a580b13a, VAE: vae-ft-mse-840000-ema-pruned.ckpt, Denoising strength: 0.3, Tiled Diffusion upscaler: SwinIR_4x, Tiled Diffusion scale factor: 1.25, Tiled Diffusion: {"Method": "Mixture of Diffusers", "Tile tile width": 96, "Tile tile height": 96, "Tile Overlap": 48, "Tile batch size": 4, "Upscaler": "SwinIR_4x", "Upscale factor": 1.25, "Keep input size": true, "NoiseInv": true, "NoiseInv Steps": 10, "NoiseInv Retouch": 1, "NoiseInv Renoise strength": 0.25, "NoiseInv Kernel size": 64}, TI hashes: "EasyNegativeV2: 339cc9210f70, negative_hand-neg: 73b524a2da12", Version: v1.6.0
postprocessing
Postprocess upscale by: 1.6, Postprocess upscaler: R-ESRGAN 4x+ Anime6B, Postprocess upscaler 2: 4x_fatal_Anime_500000_G, GFPGAN visibility: 0.35, CodeFormer visibility: 0.1, CodeFormer weight: 0.5
Extension
Postprocess upscale by: 1.6, Postprocess upscaler: R-ESRGAN 4x+ Anime6B, Postprocess upscaler 2: 4x_fatal_Anime_500000_G, GFPGAN visibility: 0.35, CodeFormer visibility: 0.1, CodeFormer weight: 0.5
Big improvement on hands with 5.0 over previous versions but hands are still the biggest weakness of this model. I know hands are hard in general but compared it with 10 models i use from time to time and this has the worst hand generation among them.
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.