V3 Image quality improvement, realism





V3 has improved the image quality【 lora :0.5-0.65】
The image has been classified into 4 categories (but it seems to have failed. You can try adding weights)
【normal,Finger insertion,nSlightly open,Bloom】
Model hash: dffc5b6e3b
VAE: sdxl_vae_fp16fix.safetensors
parameters:shangfei-yin,pussy,close-up,
Negative prompt: (pubic hair,female pubic hair),badhandv4, easynegative, FastNegativeV2, ng_deepnegative_v1_75t, extra limbs, pubic hair,sfw,full body, upper body, photo date watermark, watermark,
Steps: 30
Sampler: DPM++ 2M SDE Heun
Schedule type: Exponential
CFG scale: 7
Size: 1024x1024,768x1024,
Clip skip: 2
V3 Image quality improvement, realism
1. 轉载模型僅供學習與交流分享,其版權及最終解释權归原作者。
2. 模型原作者如需認領模型,請通過官方渠道联系海藝AI工作人員進行認證。我們致力於保護每一位創作者的權益。 點擊去認領
