HALP: Hardware-Aware Latency Pruning
This work addresses the need for efficient neural network deployment on hardware, offering a method that balances accuracy and latency for tasks like classification and detection, though it is incremental as it builds on existing pruning techniques.
The paper tackles the problem of structural pruning to improve neural network inference speed by proposing HALP, a hardware-aware latency pruning method that formulates pruning as a global resource allocation optimization, resulting in improved throughput (e.g., 1.60x for ResNet-50 on ImageNet with +0.3% accuracy change) while maintaining accuracy.
Structural pruning can simplify network architecture and improve inference speed. We propose Hardware-Aware Latency Pruning (HALP) that formulates structural pruning as a global resource allocation optimization problem, aiming at maximizing the accuracy while constraining latency under a predefined budget. For filter importance ranking, HALP leverages latency lookup table to track latency reduction potential and global saliency score to gauge accuracy drop. Both metrics can be evaluated very efficiently during pruning, allowing us to reformulate global structural pruning under a reward maximization problem given target constraint. This makes the problem solvable via our augmented knapsack solver, enabling HALP to surpass prior work in pruning efficacy and accuracy-efficiency trade-off. We examine HALP on both classification and detection tasks, over varying networks, on ImageNet and VOC datasets. In particular, for ResNet-50/-101 pruning on ImageNet, HALP improves network throughput by $1.60\times$/$1.90\times$ with $+0.3\%$/$-0.2\%$ top-1 accuracy changes, respectively. For SSD pruning on VOC, HALP improves throughput by $1.94\times$ with only a $0.56$ mAP drop. HALP consistently outperforms prior art, sometimes by large margins.