yuzu_feet174pony_lora

可以生成玉足

所使用的底层大模型是:ponyDiffusionV6

LORA推荐权重:0.8左右

正面起手词: score_9,score_8_up,score_7_up

生成参数推荐:

采样方式:Euler a

迭代步数:20到40之间

高分辨率修复开

重绘采样步数:20到30之间

重绘幅度:0.5、0.75

放大算法:R-ESRGAN 4x+ Anime6B

放大倍率:1.5、2

宽度:768

高度:1024 (64倍数的分辨率都行，最低512)

ADetailer启用 (要开)

ad第一个单元模型: mediapipe_face_full

ad第二个单元模型: face_yolov8n.pt

ad第三个单元模型: mediapipe_face_mesh_eyes_only

负面提示词: score_4,score_3,score_2,score_1,(worst quality, low quality:1.4),[:(mutated hands and fingers,one hand with more than 5 fingers, one hand with less than 5 fingers):0.8],(Multiple people),(watermark,username,artist name:1.2),lowres,bad anatomy,bad hands,text,error,missing fingers,extra digit,fewer digits,cropped,worst quality,low quality,normal quality,jpeg artifacts,signature,blurry

Can generate jade foot

The underlying large model used is: ponyDiffusionV6

Lora recommendation weight: about 0.8

Front Start Word : score_9,score_8_up,score_7_up

Generate Parameter Recommendations:

Sampling mode: Euler a

Number of iteration steps: Between 20 and 40

High Resolution Repair Development

Redraw sample steps: Between 20 and 30

Redraw amplitude: 0.5, 0.75

Amplification algorithm: R-ESRGAN 4x + Anime6B

Magnification: 1.5, 2

Width: 768

Height: 1024 (64 times resolution, 512 minimum)

ADetailer Enabled (To be On)

ad First unit model: mediapipe_face_full

ad Second unit model: face_yolov8n.pt

ad Third unit model: mediapipe_face_mesh_eyes_only

Negative : score_4,score_3,score_2,score_1, (worst quality, low quality: 1.4), [: (mutated hands and fingers,one hand with more than 5 fingers, one hand with less than 5 fingers):0.8],), (watermark,username,artist name:1.2),lowres,bad anatomy,bad hands,text,error,missing fingers,extra digit,fewer digits,cropped,worst quality,low quality,normal quality,jpeg artifacts,signature,blurry)