Discrimination-aware Network Pruning for Deep Model Compression

• URL: http://arxiv.org/abs/2001.01050v2
• Date: Mon, 29 Mar 2021 15:52:18 GMT
• Title: Discrimination-aware Network Pruning for Deep Model Compression
• Authors: Jing Liu, Bohan Zhuang, Zhuangwei Zhuang, Yong Guo, Junzhou Huang, Jinhui Zhu, Mingkui Tan
• Abstract summary: Existing pruning methods either train from scratch with sparsity constraints or minimize the reconstruction error between the feature maps of the pre-trained models and the compressed ones. We propose a simple-yet-effective method called discrimination-aware channel pruning (DCP) to choose the channels that actually contribute to the discriminative power. Experiments on both image classification and face recognition demonstrate the effectiveness of our methods.
• Score: 79.44318503847136
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: We study network pruning which aims to remove redundant channels/kernels and hence speed up the inference of deep networks. Existing pruning methods either train from scratch with sparsity constraints or minimize the reconstruction error between the feature maps of the pre-trained models and the compressed ones. Both strategies suffer from some limitations: the former kind is computationally expensive and difficult to converge, while the latter kind optimizes the reconstruction error but ignores the discriminative power of channels. In this paper, we propose a simple-yet-effective method called discrimination-aware channel pruning (DCP) to choose the channels that actually contribute to the discriminative power. Note that a channel often consists of a set of kernels. Besides the redundancy in channels, some kernels in a channel may also be redundant and fail to contribute to the discriminative power of the network, resulting in kernel level redundancy. To solve this, we propose a discrimination-aware kernel pruning (DKP) method to further compress deep networks by removing redundant kernels. To prevent DCP/DKP from selecting redundant channels/kernels, we propose a new adaptive stopping condition, which helps to automatically determine the number of selected channels/kernels and often results in more compact models with better performance. Extensive experiments on both image classification and face recognition demonstrate the effectiveness of our methods. For example, on ILSVRC-12, the resultant ResNet-50 model with 30% reduction of channels even outperforms the baseline model by 0.36% in terms of Top-1 accuracy. The pruned MobileNetV1 and MobileNetV2 achieve 1.93x and 1.42x inference acceleration on a mobile device, respectively, with negligible performance degradation. The source code and the pre-trained models are available at https://github.com/SCUT-AILab/DCP.

Related papers

• Pruning Very Deep Neural Network Channels for Efficient Inference [6.497816402045099]
  Given a trained CNN model, we propose an iterative two-step algorithm to effectively prune each layer. VGG-16 achieves the state-of-the-art results by 5x speed-up along with only 0.3% increase of error. Our method is able to accelerate modern networks like ResNet, Xception and suffers only 1.4%, 1.0% accuracy loss under 2x speed-up respectively.
  arXiv  Detail & Related papers  (2022-11-14T06:48:33Z)
• CATRO: Channel Pruning via Class-Aware Trace Ratio Optimization [61.71504948770445]
  We propose a novel channel pruning method via Class-Aware Trace Ratio Optimization (CATRO) to reduce the computational burden and accelerate the model inference. We show that CATRO achieves higher accuracy with similar cost or lower cost with similar accuracy than other state-of-the-art channel pruning algorithms. Because of its class-aware property, CATRO is suitable to prune efficient networks adaptively for various classification subtasks, enhancing handy deployment and usage of deep networks in real-world applications.
  arXiv  Detail & Related papers  (2021-10-21T06:26:31Z)
• GDP: Stabilized Neural Network Pruning via Gates with Differentiable Polarization [84.57695474130273]
  Gate-based or importance-based pruning methods aim to remove channels whose importance is smallest. GDP can be plugged before convolutional layers without bells and whistles, to control the on-and-off of each channel. Experiments conducted over CIFAR-10 and ImageNet datasets show that the proposed GDP achieves the state-of-the-art performance.
  arXiv  Detail & Related papers  (2021-09-06T03:17:10Z)
• Manipulating Identical Filter Redundancy for Efficient Pruning on Deep and Complicated CNN [126.88224745942456]
  We propose a novel Centripetal SGD (C-SGD) to make some filters identical, resulting in ideal redundancy patterns. C-SGD delivers better performance because the redundancy is better organized, compared to the existing methods.
  arXiv  Detail & Related papers  (2021-07-30T06:18:19Z)
• BWCP: Probabilistic Learning-to-Prune Channels for ConvNets via Batch Whitening [63.081808698068365]
  This work presents a probabilistic channel pruning method to accelerate Convolutional Neural Networks (CNNs) Previous pruning methods often zero out unimportant channels in training in a deterministic manner, which reduces CNN's learning capacity and results in suboptimal performance. We develop a probability-based pruning algorithm, called batch whitening channel pruning (BWCP), which canally discard unimportant channels by modeling the probability of a channel being activated.
  arXiv  Detail & Related papers  (2021-05-13T17:00:05Z)
• UCP: Uniform Channel Pruning for Deep Convolutional Neural Networks Compression and Acceleration [24.42067007684169]
  We propose a novel uniform channel pruning (UCP) method to prune deep CNN. The unimportant channels, including convolutional kernels related to them, are pruned directly. We verify our method on CIFAR-10, CIFAR-100 and ILSVRC-2012 for image classification.
  arXiv  Detail & Related papers  (2020-10-03T01:51:06Z)
• Channel Equilibrium Networks for Learning Deep Representation [63.76618960820138]
  This work shows that the combination of normalization and rectified linear function leads to inhibited channels. Unlike prior arts that simply removed the inhibited channels, we propose to "wake them up" during training by designing a novel neural building block. Channel Equilibrium (CE) block enables channels at the same layer to contribute equally to the learned representation.
  arXiv  Detail & Related papers  (2020-02-29T09:02:31Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.