Skip to content

PALP

Personalise image generation around one prompt

Visit official site

What PALP is

The paper presents a personalisation approach for text-to-image models that focuses on a single prompt. It aims to improve text alignment for complex prompts.

What you can do with it

  • Personalise a model around one prompt
  • Work with multiple subjects and reference images
  • Compare results quantitatively and qualitatively

Method notes

  • An extra score distillation sampling term is used
  • The work addresses a trade-off with prompt alignment
  • Multiple-shot and single-shot settings are demonstrated

Pros & cons

✓ What we like

  • Complex prompts can be handled with better text alignment
  • Multiple-shot and single-shot settings are shown

! What to watch out for

  • The approach is tied to one target prompt
  • Alignment depends on the added sampling term

FAQ

What is the method called?

The material calls it prompt-aligned personalisation.

What comparisons are listed?

The features list quantitative and qualitative comparisons.

Last reviewed: 2026-11-15

More AI image generation tools

View all →

How we review