This model is based on the anime AnyLoRA and incorporates personal LoRA

v1.1

Retains more versatility while giving the artwork a more stylized look (characters lean towards Ghibli, brush strokes lean towards hand-drawn style).

Suitable for animation storyboards, children's picture books, wallpapers, promotional illustrations, etc.

CFG has better effects between 5~6, high-resolution repair of the image is required.

Recommended Image

Keywords Prompt                           (Masterpiece), (Best Quality),...

Negative Keywords Negative Prompt    ng_deepnegative_v1_75t, (badhandv4:1.2), EasyNegative,

Sampling MethodSampler method        DPM++ 2M Karras,

IterationsSampling Steps          24

Image SizeWidth & Height   512*768

High-Resolution Repair Algorithm Enlargement R-ESRGAN_4x+,

Redrawing IntensityDenoising 0.4~0.6(Gradually increase based on image defects or facial blurring)

Facial blurring can also enable ADetailer, select face_yolov8n