[新手包] 窮撚玩 AI - 用 stable diffusion 砌本寫真畀自己 J

都係error

you double the "set CAMMANDLINE_ARGS"!
just replace the whole line with this

set COMMANDLINE_ARGS= --precision full --no-half --xformers

this

已知用火，個人嘗試點樣去"微"調改圖，或者係base on舊圖加新野

都係唔得

有冇方法解決

剷咗佢重新裝過

再試下

話說我一用 xformer 就會唔夠 vram, 唯有唔用
不過啲圖真係好撚癲 (Crop 咗下半幅, 廢事被 ban)


blue skirt, hyperealistic, best quality,ultra high res,(photo realistic:1.4),1 girl, iu1, adult, japanese, cinematic lighting, focus on face, (bondage:1.5), shibari, (spread legs:1.5), (bind breasts by brown rope:1.5),(bind arms behind), (bind body by rope), (tie up thighs and shins by rope:2), blindfold, (wide hips, big breasts,thick thighs),Japanese, (((school uniform))), hotel bed, ponytail, long hair, smiling, romantic style, nighttime

Negative prompt: nsfw, child, sunlight, daytime, sketches, (worst quality:2), (low quality:2), (normal quality:2), lowres, normal quality, ((monochrome)), ((grayscale)), skin spots, acnes, skin blemishes, bad anatomy,DeepNegative,(fat:1.2),bad anatomy,bad hands, text, error, missing fingers,extra digit, fewer digits, cropped, worstquality, low quality, normal quality,jpegartifacts,signature, watermark, username,blurry,bad feet,cropped,poorly drawn hands,poorly drawn face,mutation,deformed,worst quality,low quality,normal quality,jpeg artifacts,((signature)),((watermark)),extra fingers,fewer digits,extra limbs,extra arms,extra legs,malformed limbs,fused fingers,too many fingers,long neck,cross-eyed,mutated hands,polar lowres,bad body,bad proportions,gross proportions,((text)),error,missing fingers,missing arms,missing legs,extra digit, extra arms, extra leg, extra foot, nipple, teeth, muscular, closed eye, small eye

Steps: 20, Sampler: DPM++ SDE Karras, CFG scale: 7.5, Seed: 3167025073, Face restoration: CodeFormer, Size: 512x1024, Model hash: fc2511737a, Model: chilloutmix_NiPrunedFp32Fix, ControlNet Enabled: True, ControlNet Module: none, ControlNet Model: control_openpose-fp16 [9ca67cc5], ControlNet Weight: 1, ControlNet Guidance Start: 0, ControlNet Guidance End: 1, AddNet Enabled: True, AddNet Module 4: LoRA, AddNet Model 4: iu_v25(0040bc33e5f4), AddNet Weight A 4: 0.3, AddNet Weight B 4: 0.3, AddNet Module 5: LoRA, AddNet Model 5: bondageSuspension_v11(e027f683f29a), AddNet Weight A 5: 0.7, AddNet Weight B 5: 0.7

xformers 照理應該係減少用 vram

我研究咗好耐webui-user setting,我張4gb卡而家可以用到最盡, 直接出768x1024 無hires fix, 兩分幾鐘.

如果只係用一個Lora, 可以出768x1152

prompt落咗Small tits

AI 對 small tits 嘅理解同我好唔同

如果big titis會點樣？

元宇宙真係好

用呢個lora gen 極都gen 唔到佢個樣
係咪要set 其他idollikeness 做0 disable 定點先得

用 img2img inpaint

big tits with blue bra

Negative prompt: asymmetrical breast
Steps: 20, Sampler: Euler a, CFG scale: 7, Seed: 37201728, Size: 512x688, Model hash: 7234b76e42, Model: chilloutmix-Ni, Denoising strength: 0.75, Mask blur: 4

勁好玩多謝巴打教學

多謝巴打。真係好撚正

最近train緊日本地雷妹既lora搞左一個星期gen左60幾個version最慘每個version都要train 20-30分鐘

呢個應該係平時嗰個size
應該要用呢張再加big tits
但係就咁望落去都已經好假

Prompt 要落返 iu1 或者 iujie
我用果個係咁先得

有無做咩特別setup 先得或者你跟邊個教學set

試下先