MoA: Mixture-of-Attention for Subject-Context Disentanglement in Personalized Image Generation
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2024
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866913342291968000 |
|---|---|
| author | Wang, Kuan-Chieh Ostashev, Daniil Fang, Yuwei Tulyakov, Sergey Aberman, Kfir |
| author_facet | Wang, Kuan-Chieh Ostashev, Daniil Fang, Yuwei Tulyakov, Sergey Aberman, Kfir |
| contents | We introduce a new architecture for personalization of text-to-image diffusion models, coined Mixture-of-Attention (MoA). Inspired by the Mixture-of-Experts mechanism utilized in large language models (LLMs), MoA distributes the generation workload between two attention pathways: a personalized branch and a non-personalized prior branch. MoA is designed to retain the original model's prior by fixing its attention layers in the prior branch, while minimally intervening in the generation process with the personalized branch that learns to embed subjects in the layout and context generated by the prior branch. A novel routing mechanism manages the distribution of pixels in each layer across these branches to optimize the blend of personalized and generic content creation. Once trained, MoA facilitates the creation of high-quality, personalized images featuring multiple subjects with compositions and interactions as diverse as those generated by the original model. Crucially, MoA enhances the distinction between the model's pre-existing capability and the newly augmented personalized intervention, thereby offering a more disentangled subject-context control that was previously unattainable. Project page: https://snap-research.github.io/mixture-of-attention |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_11565 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | MoA: Mixture-of-Attention for Subject-Context Disentanglement in Personalized Image Generation Wang, Kuan-Chieh Ostashev, Daniil Fang, Yuwei Tulyakov, Sergey Aberman, Kfir Computer Vision and Pattern Recognition Artificial Intelligence Graphics We introduce a new architecture for personalization of text-to-image diffusion models, coined Mixture-of-Attention (MoA). Inspired by the Mixture-of-Experts mechanism utilized in large language models (LLMs), MoA distributes the generation workload between two attention pathways: a personalized branch and a non-personalized prior branch. MoA is designed to retain the original model's prior by fixing its attention layers in the prior branch, while minimally intervening in the generation process with the personalized branch that learns to embed subjects in the layout and context generated by the prior branch. A novel routing mechanism manages the distribution of pixels in each layer across these branches to optimize the blend of personalized and generic content creation. Once trained, MoA facilitates the creation of high-quality, personalized images featuring multiple subjects with compositions and interactions as diverse as those generated by the original model. Crucially, MoA enhances the distinction between the model's pre-existing capability and the newly augmented personalized intervention, thereby offering a more disentangled subject-context control that was previously unattainable. Project page: https://snap-research.github.io/mixture-of-attention |
| title | MoA: Mixture-of-Attention for Subject-Context Disentanglement in Personalized Image Generation |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence Graphics |
| url | https://arxiv.org/abs/2404.11565 |