LSGQuant: Layer-Sensitivity Guided Quantization for One-Step Diffusion Real-World Video Super-Resolution
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866917244449062912 |
|---|---|
| author | Wu, Tianxing Chen, Zheng Xu, Cirou Chai, Bowen Guo, Yong Liu, Yutong Kong, Linghe Zhang, Yulun |
| author_facet | Wu, Tianxing Chen, Zheng Xu, Cirou Chai, Bowen Guo, Yong Liu, Yutong Kong, Linghe Zhang, Yulun |
| contents | One-Step Diffusion Models have demonstrated promising capability and fast inference in video super-resolution (VSR) for real-world. Nevertheless, the substantial model size and high computational cost of Diffusion Transformers (DiTs) limit downstream applications. While low-bit quantization is a common approach for model compression, the effectiveness of quantized models is challenged by the high dynamic range of input latent and diverse layer behaviors. To deal with these challenges, we introduce LSGQuant, a layer-sensitivity guided quantizing approach for one-step diffusion-based real-world VSR. Our method incorporates a Dynamic Range Adaptive Quantizer (DRAQ) to fit video token activations. Furthermore, we estimate layer sensitivity and implement a Variance-Oriented Layer Training Strategy (VOLTS) by analyzing layer-wise statistics in calibration. We also introduce Quantization-Aware Optimization (QAO) to jointly refine the quantized branch and a retained high-precision branch. Extensive experiments demonstrate that our method has nearly performance to origin model with full-precision and significantly exceeds existing quantization techniques. Code is available at: https://github.com/zhengchen1999/LSGQuant. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_03182 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | LSGQuant: Layer-Sensitivity Guided Quantization for One-Step Diffusion Real-World Video Super-Resolution Wu, Tianxing Chen, Zheng Xu, Cirou Chai, Bowen Guo, Yong Liu, Yutong Kong, Linghe Zhang, Yulun Computer Vision and Pattern Recognition One-Step Diffusion Models have demonstrated promising capability and fast inference in video super-resolution (VSR) for real-world. Nevertheless, the substantial model size and high computational cost of Diffusion Transformers (DiTs) limit downstream applications. While low-bit quantization is a common approach for model compression, the effectiveness of quantized models is challenged by the high dynamic range of input latent and diverse layer behaviors. To deal with these challenges, we introduce LSGQuant, a layer-sensitivity guided quantizing approach for one-step diffusion-based real-world VSR. Our method incorporates a Dynamic Range Adaptive Quantizer (DRAQ) to fit video token activations. Furthermore, we estimate layer sensitivity and implement a Variance-Oriented Layer Training Strategy (VOLTS) by analyzing layer-wise statistics in calibration. We also introduce Quantization-Aware Optimization (QAO) to jointly refine the quantized branch and a retained high-precision branch. Extensive experiments demonstrate that our method has nearly performance to origin model with full-precision and significantly exceeds existing quantization techniques. Code is available at: https://github.com/zhengchen1999/LSGQuant. |
| title | LSGQuant: Layer-Sensitivity Guided Quantization for One-Step Diffusion Real-World Video Super-Resolution |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2602.03182 |