QD-BEV : Quantization-aware View-guided Distillation for Multi-view 3D Object Detection

TL;DR

This paper introduces QD-BEV, a quantization-aware view-guided distillation method for multi-view 3D object detection that stabilizes training and maintains high accuracy with significantly compressed models.

Contribution

The paper proposes a novel view-guided distillation objective to improve quantization-aware training for BEV-based 3D detection models, enabling efficient deployment.

Findings

Achieves comparable or better accuracy than prior methods with high model compression.

On nuScenes, the 4-bit weight and 6-bit activation QD-BEV-Tiny reaches 37.2% NDS, outperforming BevFormer-Tiny.

Models demonstrate excellent performance across different sizes with substantial efficiency gains.

Abstract

Multi-view 3D detection based on BEV (bird-eye-view) has recently achieved significant improvements. However, the huge memory consumption of state-of-the-art models makes it hard to deploy them on vehicles, and the non-trivial latency will affect the real-time perception of streaming applications. Despite the wide application of quantization to lighten models, we show in our paper that directly applying quantization in BEV tasks will 1) make the training unstable, and 2) lead to intolerable performance degradation. To solve these issues, our method QD-BEV enables a novel view-guided distillation (VGD) objective, which can stabilize the quantization-aware training (QAT) while enhancing the model performance by leveraging both image features and BEV features. Our experiments show that QD-BEV achieves similar or even better accuracy than previous methods with significant efficiency gains. On the nuScenes datasets, the 4-bit weight and 6-bit activation quantized QD-BEV-Tiny model achieves 37.2% NDS with only 15.8 MB model size, outperforming BevFormer-Tiny by 1.8% with an 8x model compression. On the Small and Base variants, QD-BEV models also perform superbly and achieve 47.9% NDS (28.2 MB) and 50.9% NDS (32.9 MB), respectively.