STARFlow: Scaling Latent Normalizing Flows for High-resolution Image Synthesis
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912418755510272 |
|---|---|
| author | Gu, Jiatao Chen, Tianrong Berthelot, David Zheng, Huangjie Wang, Yuyang Zhang, Ruixiang Dinh, Laurent Bautista, Miguel Angel Susskind, Josh Zhai, Shuangfei |
| author_facet | Gu, Jiatao Chen, Tianrong Berthelot, David Zheng, Huangjie Wang, Yuyang Zhang, Ruixiang Dinh, Laurent Bautista, Miguel Angel Susskind, Josh Zhai, Shuangfei |
| contents | We present STARFlow, a scalable generative model based on normalizing flows that achieves strong performance in high-resolution image synthesis. The core of STARFlow is Transformer Autoregressive Flow (TARFlow), which combines the expressive power of normalizing flows with the structured modeling capabilities of Autoregressive Transformers. We first establish the theoretical universality of TARFlow for modeling continuous distributions. Building on this foundation, we introduce several key architectural and algorithmic innovations to significantly enhance scalability: (1) a deep-shallow design, wherein a deep Transformer block captures most of the model representational capacity, complemented by a few shallow Transformer blocks that are computationally efficient yet substantially beneficial; (2) modeling in the latent space of pretrained autoencoders, which proves more effective than direct pixel-level modeling; and (3) a novel guidance algorithm that significantly boosts sample quality. Crucially, our model remains an end-to-end normalizing flow, enabling exact maximum likelihood training in continuous spaces without discretization. STARFlow achieves competitive performance in both class-conditional and text-conditional image generation tasks, approaching state-of-the-art diffusion models in sample quality. To our knowledge, this work is the first successful demonstration of normalizing flows operating effectively at this scale and resolution. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_06276 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | STARFlow: Scaling Latent Normalizing Flows for High-resolution Image Synthesis Gu, Jiatao Chen, Tianrong Berthelot, David Zheng, Huangjie Wang, Yuyang Zhang, Ruixiang Dinh, Laurent Bautista, Miguel Angel Susskind, Josh Zhai, Shuangfei Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning We present STARFlow, a scalable generative model based on normalizing flows that achieves strong performance in high-resolution image synthesis. The core of STARFlow is Transformer Autoregressive Flow (TARFlow), which combines the expressive power of normalizing flows with the structured modeling capabilities of Autoregressive Transformers. We first establish the theoretical universality of TARFlow for modeling continuous distributions. Building on this foundation, we introduce several key architectural and algorithmic innovations to significantly enhance scalability: (1) a deep-shallow design, wherein a deep Transformer block captures most of the model representational capacity, complemented by a few shallow Transformer blocks that are computationally efficient yet substantially beneficial; (2) modeling in the latent space of pretrained autoencoders, which proves more effective than direct pixel-level modeling; and (3) a novel guidance algorithm that significantly boosts sample quality. Crucially, our model remains an end-to-end normalizing flow, enabling exact maximum likelihood training in continuous spaces without discretization. STARFlow achieves competitive performance in both class-conditional and text-conditional image generation tasks, approaching state-of-the-art diffusion models in sample quality. To our knowledge, this work is the first successful demonstration of normalizing flows operating effectively at this scale and resolution. |
| title | STARFlow: Scaling Latent Normalizing Flows for High-resolution Image Synthesis |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2506.06276 |