SANA 1.5: Efficient Scaling of Training-Time and Inference-Time Compute in Linear Diffusion Transformer

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Xie, Enze, Chen, Junsong, Zhao, Yuyang, Yu, Jincheng, Zhu, Ligeng, Wu, Chengyue, Lin, Yujun, Zhang, Zhekai, Li, Muyang, Chen, Junyu, Cai, Han, Liu, Bingchen, Zhou, Daquan, Han, Song
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866916741211226112
author Xie, Enze
Chen, Junsong
Zhao, Yuyang
Yu, Jincheng
Zhu, Ligeng
Wu, Chengyue
Lin, Yujun
Zhang, Zhekai
Li, Muyang
Chen, Junyu
Cai, Han
Liu, Bingchen
Zhou, Daquan
Han, Song
author_facet Xie, Enze
Chen, Junsong
Zhao, Yuyang
Yu, Jincheng
Zhu, Ligeng
Wu, Chengyue
Lin, Yujun
Zhang, Zhekai
Li, Muyang
Chen, Junyu
Cai, Han
Liu, Bingchen
Zhou, Daquan
Han, Song
contents This paper presents SANA-1.5, a linear Diffusion Transformer for efficient scaling in text-to-image generation. Building upon SANA-1.0, we introduce three key innovations: (1) Efficient Training Scaling: A depth-growth paradigm that enables scaling from 1.6B to 4.8B parameters with significantly reduced computational resources, combined with a memory-efficient 8-bit optimizer. (2) Model Depth Pruning: A block importance analysis technique for efficient model compression to arbitrary sizes with minimal quality loss. (3) Inference-time Scaling: A repeated sampling strategy that trades computation for model capacity, enabling smaller models to match larger model quality at inference time. Through these strategies, SANA-1.5 achieves a text-image alignment score of 0.81 on GenEval, which can be further improved to 0.96 through inference scaling with VILA-Judge, establishing a new SoTA on GenEval benchmark. These innovations enable efficient model scaling across different compute budgets while maintaining high quality, making high-quality image generation more accessible. Our code and pre-trained models are released.
format Preprint
id arxiv_https___arxiv_org_abs_2501_18427
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle SANA 1.5: Efficient Scaling of Training-Time and Inference-Time Compute in Linear Diffusion Transformer
Xie, Enze
Chen, Junsong
Zhao, Yuyang
Yu, Jincheng
Zhu, Ligeng
Wu, Chengyue
Lin, Yujun
Zhang, Zhekai
Li, Muyang
Chen, Junyu
Cai, Han
Liu, Bingchen
Zhou, Daquan
Han, Song
Computer Vision and Pattern Recognition
This paper presents SANA-1.5, a linear Diffusion Transformer for efficient scaling in text-to-image generation. Building upon SANA-1.0, we introduce three key innovations: (1) Efficient Training Scaling: A depth-growth paradigm that enables scaling from 1.6B to 4.8B parameters with significantly reduced computational resources, combined with a memory-efficient 8-bit optimizer. (2) Model Depth Pruning: A block importance analysis technique for efficient model compression to arbitrary sizes with minimal quality loss. (3) Inference-time Scaling: A repeated sampling strategy that trades computation for model capacity, enabling smaller models to match larger model quality at inference time. Through these strategies, SANA-1.5 achieves a text-image alignment score of 0.81 on GenEval, which can be further improved to 0.96 through inference scaling with VILA-Judge, establishing a new SoTA on GenEval benchmark. These innovations enable efficient model scaling across different compute budgets while maintaining high quality, making high-quality image generation more accessible. Our code and pre-trained models are released.
title SANA 1.5: Efficient Scaling of Training-Time and Inference-Time Compute in Linear Diffusion Transformer
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2501.18427