QuantFace: Efficient Quantization for Face Restoration

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Jiatong, Zhu, Libo, Qin, Haotong, Wang, Jingkai, Kong, Linghe, Chen, Guihai, Zhang, Yulun, Yang, Xiaokang
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918212134764544
author Li, Jiatong
Zhu, Libo
Qin, Haotong
Wang, Jingkai
Kong, Linghe
Chen, Guihai
Zhang, Yulun
Yang, Xiaokang
author_facet Li, Jiatong
Zhu, Libo
Qin, Haotong
Wang, Jingkai
Kong, Linghe
Chen, Guihai
Zhang, Yulun
Yang, Xiaokang
contents Diffusion models have been achieving remarkable performance in face restoration. However, the heavy computations hamper the widespread adoption of these models. In this work, we propose QuantFace, a novel low-bit quantization framework for face restoration models, where the full-precision (i.e., 32-bit) weights and activations are quantized to 4~6-bit. We first analyze the data distribution within activations and find that it is highly variant. To preserve the original data information, we employ rotation-scaling channel balancing. Furthermore, we propose Quantization-Distillation Low-Rank Adaptation (QD-LoRA), which jointly optimizes for quantization and distillation performance. Finally, we propose an adaptive bit-width allocation strategy. We formulate such a strategy as an integer programming problem that combines quantization error and perceptual metrics to find a satisfactory resource allocation. Extensive experiments on the synthetic and real-world datasets demonstrate the effectiveness of QuantFace under 6-bit and 4-bit. QuantFace achieves significant advantages over recent leading low-bit quantization methods for face restoration. The code is available at https://github.com/jiatongli2024/QuantFace.
format Preprint
id arxiv_https___arxiv_org_abs_2506_00820
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle QuantFace: Efficient Quantization for Face Restoration
Li, Jiatong
Zhu, Libo
Qin, Haotong
Wang, Jingkai
Kong, Linghe
Chen, Guihai
Zhang, Yulun
Yang, Xiaokang
Computer Vision and Pattern Recognition
Diffusion models have been achieving remarkable performance in face restoration. However, the heavy computations hamper the widespread adoption of these models. In this work, we propose QuantFace, a novel low-bit quantization framework for face restoration models, where the full-precision (i.e., 32-bit) weights and activations are quantized to 4~6-bit. We first analyze the data distribution within activations and find that it is highly variant. To preserve the original data information, we employ rotation-scaling channel balancing. Furthermore, we propose Quantization-Distillation Low-Rank Adaptation (QD-LoRA), which jointly optimizes for quantization and distillation performance. Finally, we propose an adaptive bit-width allocation strategy. We formulate such a strategy as an integer programming problem that combines quantization error and perceptual metrics to find a satisfactory resource allocation. Extensive experiments on the synthetic and real-world datasets demonstrate the effectiveness of QuantFace under 6-bit and 4-bit. QuantFace achieves significant advantages over recent leading low-bit quantization methods for face restoration. The code is available at https://github.com/jiatongli2024/QuantFace.
title QuantFace: Efficient Quantization for Face Restoration
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2506.00820