Task-Specific Zero-shot Quantization-Aware Training for Object Detection

TL;DR

This paper introduces a task-specific zero-shot quantization-aware training method for object detection that synthesizes calibration data and uses knowledge distillation to improve quantized network performance without real training data.

Contribution

It proposes a novel framework combining task-specific data synthesis and knowledge distillation for zero-shot quantization in object detection.

Findings

Achieves state-of-the-art performance on MS-COCO and Pascal VOC datasets.

Effectively reconstructs object locations, sizes, and categories without prior knowledge.

Demonstrates efficiency and robustness of the proposed method.

Abstract

Quantization is a key technique to reduce network size and computational complexity by representing the network parameters with a lower precision. Traditional quantization methods rely on access to original training data, which is often restricted due to privacy concerns or security challenges. Zero-shot Quantization (ZSQ) addresses this by using synthetic data generated from pre-trained models, eliminating the need for real training data. Recently, ZSQ has been extended to object detection. However, existing methods use unlabeled task-agnostic synthetic images that lack the specific information required for object detection, leading to suboptimal performance. In this paper, we propose a novel task-specific ZSQ framework for object detection networks, which consists of two main stages. First, we introduce a bounding box and category sampling strategy to synthesize a task-specific calibration set from the pre-trained network, reconstructing object locations, sizes, and category distributions without any prior knowledge. Second, we integrate task-specific training into the knowledge distillation process to restore the performance of quantized detection networks. Extensive experiments conducted on the MS-COCO and Pascal VOC datasets demonstrate the efficiency and state-of-the-art performance of our method. Our code is publicly available at: https://github.com/DFQ-Dojo/dfq-toolkit .