Models>
XingQiu v2.0 | Genshin
Quick Generate
Model Description by Creator
My pixiv link:
https://www.pixiv.net/users/76094070
updated v2.0
This time I used locon and offset noise, got another effect, if you do not like this effect, please download v1.0, LORA weighting, my recommendation is 0.5 ~ 0.8, here are some control charts.
↓↓↓↓here is the v1.0_step_effect drawing↓↓↓↓
<lora:xingqiu:0.5>,masterpiece,best quality,official art,extremely detailed CG unity 8k wallpaper,1boy,solo,seductive smile, sitting_chair, ((masterpiece,best quality)), dynamic angle, (beautiful delicate face:1.1), solo,(1boy:1.3), masterpiece, best quality,
Negative prompt: EasyNegative,(((simple background))),monochrome ,lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry, lowres, bad anatomy, bad hands, text, error, extra digit, fewer digits, cropped, (breasts),worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry, ugly,pregnant,vore,duplicate,morbid,mutilated,tran nsexual, hermaphrodite,long neck,mutated hands,poorly drawn hands,poorly drawn face,mutation,deformed,blurry,bad anatomy,bad proportions,malformed limbs,extra limbs,cloned face,disfigured,gross proportions, (((missing arms))),(((missing legs))), (((extra arms))),(((extra legs))),pubic hair, plump,bad legs,error legs,username,blurry,bad feet,, lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry
Steps: 20, Sampler: DPM++ SDE Karras, CFG scale: 7, Seed: 1363376316, Size: 512x768, Model hash: e4b17ce185, Clip skip: 2, ENSD: 31337
Sorry about my horrible English. GOOGLE translate： I'm just a novice player, just follow the tutorial and don't understand the logic and principle of training. why no one likes little boys HERE IS XINGQIU. use 32 pic to train this lora max_train_epoches:40 Then this is the result of my 18th test and found that the effect is not bad. I plan to train for 40 reps, this is the 18th version ↓↓↓↓Here is the v0.18_step_effect drawing↓↓↓↓
* represents the value of the variable<lora:xingqiu-000018:*>,((masterpiece,best quality)), dynamic angle, (beautiful delicate face:1.1),1boy,solo, light blue hair, fluffy medium-short hair, bangs,earrings, brown eyes, blue and black Military highschool dress,((school uniform))
Negative prompt: EasyNegative,(((simple background))),worst fingers
Steps: *, Sampler: DPM++ 2S a Karras, CFG scale: 8, Seed: 1784626632, Size: 512x768, Model hash: fc2511737a, Model: anything-v4.5-pruned.safetensors, Clip skip: 2
Used embeddings: EasyNegative [119b]
<lora:xingqiu-000018:*>,((masterpiece,best quality)), dynamic angle, (beautiful delicate face:1.1),1boy,solo, light blue hair, fluffy medium-short hair, bangs,earrings, brown eyes, blue and black Military highschool dress,((school uniform))
Negative prompt: EasyNegative,(((simple background))),worst fingers
Steps: *, Sampler: DPM++ 2S a Karras, CFG scale: 8, Seed: 1784626632, Size: 512x768, Model hash: fc2511737a, Model: chilloutmix_NiPrunedFp32Fix, Clip skip: 2
Used embeddings: EasyNegative [119b]
