Gradio

Nested Attention: Semantic-aware Attention Values for Concept Personalization

Upload up to 3 reference images. Faces will be auto-aligned before personalization. Include the placeholder token (e.g., <person>) in your prompt, set token weight, and choose how many outputs you want.

Reference Image 1

Reference Image 2 (optional)

Reference Image 3 (optional)

Prompt

Special Token Weight (w)

1 5

Number of Images to Generate

1 6

Random Seed (use -1 for random and up to 100000)

-1 100000

Example Prompts

Reference Image 1	Prompt	Special Token Weight (w)	Number of Images to Generate	Random Seed (use -1 for random and up to 100000)

Generated Images