Exploring Linear Relationship in Feature Map Subspace for ConvNets Compression

TL;DR

This paper introduces a novel filter pruning method for CNNs based on the linear relationships in feature map subspaces, effectively reducing redundancy and improving model efficiency without structural constraints.

Contribution

It proposes a subspace clustering approach to filter pruning that leverages linear relationships in feature maps, outperforming existing methods and being adaptable to any network architecture.

Findings

Outperforms existing pruning techniques before fine-tuning

Achieves state-of-the-art results after fine-tuning

Independent of network structure, applicable to various models

Abstract

While the research on convolutional neural networks (CNNs) is progressing quickly, the real-world deployment of these models is often limited by computing resources and memory constraints. In this paper, we address this issue by proposing a novel filter pruning method to compress and accelerate CNNs. Our work is based on the linear relationship identified in different feature map subspaces via visualization of feature maps. Such linear relationship implies that the information in CNNs is redundant. Our method eliminates the redundancy in convolutional filters by applying subspace clustering to feature maps. In this way, most of the representative information in the network can be retained in each cluster. Therefore, our method provides an effective solution to filter pruning for which most existing methods directly remove filters based on simple heuristics. The proposed method is independent of the network structure, thus it can be adopted by any off-the-shelf deep learning libraries. Experiments on different networks and tasks show that our method outperforms existing techniques before fine-tuning, and achieves the state-of-the-art results after fine-tuning.