paper-with-me

Papers

TQ-DiT: Efficient Time-Aware Quantization for Diffusion Transformers

2025-02-06 · Younghye Hwang, Hyojin Lee, Joonhyuk Kang

Diffusion transformers (DiTs) combine transformer architectures with diffusion models. However, their computational complexity imposes significant limitations on real-time applications and sustainability of AI systems. In this study, we aim to enhance the computational efficiency through model quantization, which represents the weights and activation values with lower precision. Multi-region quantization (MRQ) is introduced to address the asymmetric distribution of network values in DiT blocks by allocating two scaling parameters to sub-regions. Additionally, time-grouping quantization (TGQ) is proposed to reduce quantization error caused by temporal variation in activations. The experimental results show that the proposed algorithm achieves performance comparable to the original full-precision model with only a 0.29 increase in FID at W8A8. Furthermore, it outperforms other baselines at W6A6, thereby confirming its suitability for low-bit quantization. These results highlight the potential of our method to enable efficient real-time generative models.

📄 PDF Abstract BibTeX arXiv:2502.04056

Code (0)

등록된 구현이 없습니다.

Tasks

Computational EfficiencyQuantization

Methods 이 논문이 사용한 방법론

Diffusion Diffusion models generate samples by gradually removing noise from a signal, and their training objective can be expressed as a reweighted variational lower-bound…

Similar Papers 제목 키워드 기반

TaQ-DiT: Time-aware Quantization for Diffusion Transformers

2024-11-21 · Xinyan Liu, Huihong Shi, Yang Xu, Zhongfeng Wang

Transformer-based diffusion models, dubbed Diffusion Transformers (DiTs), have achieved state-of-the-art performance in image and video generation tasks. However, their large model size and slow inference speed limit the…

DenoisingModel CompressionQuantizationVideo Generation

Effective Quantization for Diffusion Models on CPUs

2023-11-02 · Hanwen Chang, Haihao Shen, Yiyang Cai, Xinyu Ye 외

Diffusion models have gained popularity for generating images from textual descriptions. Nonetheless, the substantial need for computational resources continues to present a noteworthy challenge, contributing to time-con…

Quantization

SemanticDialect: Semantic-Aware Mixed-Format Quantization for Video Diffusion Transformers

2026-03-03 · Wonsuk Jang, Thierry Tambe arxiv

Diffusion Transformers (DiTs) achieve state-of-the-art video generation quality, but their substantial memory and computational footprints hinder edge deployment. Quantization can reduce these costs, yet existing methods…

Video Generation

ViDiT-Q: Efficient and Accurate Quantization of Diffusion Transformers for Image and Video Generation

2024-06-04 · Tianchen Zhao, Tongcheng Fang, Enshu Liu, Rui Wan 외

Diffusion transformers (DiTs) have exhibited remarkable performance in visual generation tasks, such as generating realistic images or videos based on textual instructions. However, larger model sizes and multi-frame pro…

QuantizationVideo Generation

ConvRot: Rotation-Based Plug-and-Play 4-bit Quantization for Diffusion Transformers

2025-12-03 · Feice Huang, Zuliang Han, Xing Zhou, Yihuang Chen 외 arxiv

Diffusion transformers have demonstrated strong capabilities in generating high-quality images. However, as model size increases, the growing memory footprint and inference latency pose significant challenges for practic…