Compressing Multi-Task Model for Autonomous Driving via Pruning and Knowledge Distillation
This addresses the problem of model size and complexity for real-time deployment on on-board devices in autonomous driving, representing an incremental improvement in compression techniques.
The paper tackled the challenge of deploying large multi-task models for autonomous driving by proposing a compression framework combining pruning and knowledge distillation, achieving a 32.7% reduction in parameters with minimal performance loss (e.g., -1.2% Recall, -1.8% mAP50) and real-time operation at 32.7 FPS.
Autonomous driving systems rely on panoptic perception to jointly handle object detection, drivable area segmentation, and lane line segmentation. Although multi-task learning is an effective way to integrate these tasks, its increasing model parameters and complexity make deployment on on-board devices difficult. To address this challenge, we propose a multi-task model compression framework that combines task-aware safe pruning with feature-level knowledge distillation. Our safe pruning strategy integrates Taylor-based channel importance with gradient conflict penalty to keep important channels while removing redundant and conflicting channels. To mitigate performance degradation after pruning, we further design a task head-agnostic distillation method that transfers intermediate backbone and encoder features from a teacher to a student model as guidance. Experiments on the BDD100K dataset demonstrate that our compressed model achieves a 32.7% reduction in parameters while segmentation performance shows negligible accuracy loss and only a minor decrease in detection (-1.2% for Recall and -1.8% for mAP50) compared to the teacher. The compressed model still runs at 32.7 FPS in real-time. These results show that combining pruning and knowledge distillation provides an effective compression solution for multi-task panoptic perception.