Magnitude Attention-based Dynamic Pruning

초록

Existing pruning methods often rely on weight importance to identify sparse structures but typically apply this information statically, without leveraging it adaptively during training. In this work, we propose a novel approach- Magnitude Attention-based Dynamic Pruning (MAP) method, which applies the importance of weights throughout both the forward and backward paths to explore sparse model structures dynamically. Magnitude attention is defined based on the magnitude of weights as continuous real-valued numbers enabling a seamless transition from a redundant to an effective sparse network by promoting efficient exploration. Additionally, the attention mechanism ensures more effective updates for important layers within the sparse network. Later, our approach shifts from exploration to exploitation, exclusively updating the sparse model composed of crucial weights based on the explored structure, resulting in pruned models that not only achieve performance comparable to dense models but also outperform previous pruning methods on CIFAR-10/100 and ImageNet.

키워드