Search

Pruning methods to compress and accelerate deep convolutional neural networks (CNNs) have recently attracted growing attention, with the view of deploying pruned networks on resource-constrained hardware devices. However, most existing methods focus on small granularities, such as weight, kernel and filter, for the exploration of pruning. Thus, it will be bound to iteratively prune the whole neural networks based on those small granularities for high compression ratio with little performance loss. To address these issues, we theoretically analyze the relationship between the activation and gradient sparsity, and the channel saliency. Based on our findings, we propose a novel and effective method of weak sub-network pruning (WSP). Specifically, for a well-trained network model, we divide the whole compression process into two non-iterative stages. The first stage is to directly obtain a strong sub-network by pruning the weakest one. We first identify the less important channels from all the layers and determine the weakest sub-network, whereby each selected channel makes a minimal contribution to both the feed-forward and feed-backward processes. Then, a one-shot pruning strategy is executed to form a strong sub-network enabling fine tuning, while significantly reducing the impact of the network depth and width on the compression efficiency, especially for deep and wide network architectures. The second stage is to globally fine-tune the strong sub-network using several epochs to restore its original recognition accuracy. Furthermore, our proposed method impacts on the fully-connected layers as well as the convolutional layers for simultaneous compression and acceleration. Comprehensive experiments on VGG16 and ResNet-50 involving a variety of popular benchmarks, such as ImageNet-1K, CIFAR-10, CUB-200 and PASCAL VOC, demonstrate that our WSP method achieves superior performance on classification, domain adaption and object detection tasks with small model size. Our source code is available at https://github.com/QingbeiGuo/WSP.git. [ABSTRACT FROM AUTHOR]

Although group convolution operators are increasingly used in deep convolutional neural networks to improve the computational efficiency and to reduce the number of parameters, most existing methods construct their group convolution architectures by a predefined partitioning of the filters of each convolutional layer into multiple regular filter groups with an equal spatial group size and data-independence, which prevents a full exploitation of their potential. To tackle this issue, we propose a novel method of designing self-grouping convolutional neural networks, called SG-CNN, in which the filters of each convolutional layer group themselves based on the similarity of their importance vectors. Concretely, for each filter, we first evaluate the importance value of their input channels to identify the importance vectors, and then group these vectors by clustering. Using the resulting data-dependent centroids, we prune the less important connections, which implicitly minimizes the accuracy loss of the pruning, thus yielding a set of diverse group convolution filters. Subsequently, we develop two fine-tuning schemes, i.e. (1) both local and global fine-tuning and (2) global only fine-tuning, which experimentally deliver comparable results, to recover the recognition capacity of the pruned network. Comprehensive experiments carried out on the CIFAR-10/100 and ImageNet datasets demonstrate that our self-grouping convolution method adapts to various state-of-the-art CNN architectures, such as ResNet and DenseNet, and delivers superior performance in terms of compression ratio, speedup and recognition accuracy. We demonstrate the ability of SG-CNN to generalize by transfer learning, including domain adaption and object detection, showing competitive results. Our source code is available at https://github.com/QingbeiGuo/SG-CNN.git. [ABSTRACT FROM AUTHOR]

• We build a new standalone control module based on the scaled sigmoid function to enrich the neural network module family to enable the neural architecture optimization. • Our DNAL method produces no candidate neural architectures but one, thus drastically improving the learning efficiency, i.e., costing 20 epochs for CIFAR-10 and 10 epochs for ImageNet. • It is applicable to conventional CNNs, lightweight CNNs, and stochastic supernets. • Extensive experiments confirm that our DNAL method achieves excellent performance on various CNN architectures, including VGG16, ResNet50, MobileNetV2, and ProxylessNAS, over the task of CIFAR-10 and ImageNet-1K classification. Efficient neural networks has received ever-increasing attention with the evolution of convolutional neural networks (CNNs), especially involving their deployment on embedded and mobile platforms. One of the biggest problems to obtaining such efficient neural networks is efficiency, even recent differentiable neural architecture search (DNAS) requires to sample a small number of candidate neural architectures for the selection of the optimal neural architecture. To address this computational efficiency issue, we introduce a novel architecture parameterization based on scaled sigmoid function , and propose a general Differentiable Neural Architecture Learning (DNAL) method to obtain efficient neural networks without the need to evaluate candidate neural networks. Specifically, for stochastic supernets as well as conventional CNNs, we build a new channel-wise module layer with the architecture components controlled by a scaled sigmoid function. We train these neural network models from scratch. The network optimization is decoupled into the weight optimization and the architecture optimization, which avoids the interaction between the two types of parameters and alleviates the vanishing gradient problem. We address the non-convex optimization problem of efficient neural networks by the continuous scaled sigmoid method instead of the common softmax method. Extensive experiments demonstrate our DNAL method delivers superior performance in terms of efficiency, and adapts to conventional CNNs (e.g., VGG16 and ResNet50), lightweight CNNs (e.g., MobileNetV2) and stochastic supernets (e.g., ProxylessNAS). The optimal neural networks learned by DNAL surpass those produced by the state-of-the-art methods on the benchmark CIFAR-10 and ImageNet-1K dataset in accuracy, model size and computational complexity. Our source code is available at https://github.com/QingbeiGuo/DNAL.git. [ABSTRACT FROM AUTHOR]