Analysis of Classifier-Free Guidance Weight Schedulers

Classifier-Free Guidance (CFG) enhances the quality and condition adherence of text-toimage diffusion models. It operates by combining the conditional and unconditional predictions using a fixed weight. However, recent works vary the weights throughout the diffusion process, reporting superior resul...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Transactions on Machine Learning Research Journal
Hlavní autoři: Wang, Xi, Dufour, Nicolas, Andreou, Nefeli, Cani, Marie-Paule, Fernández Abrevaya, Victoria, Picard, David, Kalogeiton, Vicky
Médium: Journal Article
Jazyk:angličtina
Vydáno: [Amherst Massachusetts]: OpenReview.net, 2022 2024
Témata:
Artificial Intelligence
Computer Science
Generative Model
Diffusion Model
ISSN:2835-8856
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Classifier-Free Guidance (CFG) enhances the quality and condition adherence of text-toimage diffusion models. It operates by combining the conditional and unconditional predictions using a fixed weight. However, recent works vary the weights throughout the diffusion process, reporting superior results but without providing any rationale or analysis. By conducting comprehensive experiments, this paper provides insights into CFG weight schedulers. Our findings suggest that simple, monotonically increasing weight schedulers consistently lead to improved performances, requiring merely a single line of code. In addition, more complex parametrized schedulers can be optimized for further improvement, but do not generalize across different models and tasks.
ISSN:2835-8856
DOI:10.48550/arXiv.2404.13040