Structured weight pruning is a representative model compression technique of DNNs for hardware efficiency and inference accelerations. Previous works in this area leave great space for improvement since sparse structures with combinations of different structured pruning schemes are not exploited fully and efficiently. To mitigate the limitations, we propose SS-Auto, a single-shot, automatic structured pruning framework that can achieve row pruning and column pruning simultaneously. We adopt soft constraint-based formulation to alleviate the strong non-convexity of l0-norm constraints used in state-of-the-art ADMM-based methods for faster convergence and fewer hyperparameters. Instead of solving the problem directly, a Primal-Proximal solution is proposed to avoid the pitfall of penalizing all weights equally, thereby enhancing the accuracy. Extensive experiments on CIFAR-10 and CIFAR-100 datasets demonstrate that the proposed framework can achieve ultra-high pruning rates while maintaining accuracy. Furthermore, significant inference speedup has been observed from the proposed framework through actual measurements on the smartphone.
We present a filter pruning approach for deep model compression, using a multitask network. Our approach is based on learning a a pruner network to prune a pre-trained target network. The pruner is essentially a multitask deep neural network with binary outputs that help identify the filters from each layer of the original network that do not have any significant contribution to the model and can therefore be pruned. The pruner network has the same architecture as the original network except that it has a multitask/multi-output last layer containing binary-valued outputs (one per filter), which indicate which filters have to be pruned. The pruner's goal is to minimize the number of filters from the original network by assigning zero weights to the corresponding output feature-maps. In contrast to most of the existing methods, instead of relying on iterative pruning, our approach can prune the network (original network) in one go and, moreover, does not require specifying the degree of pruning for each layer (and can learn it instead). The compressed model produced by our approach is generic and does not need any special hardware/software support. Moreover, augmenting with other methods such as knowledge distillation, quantization, and connection pruning can increase the degree of compression for the proposed approach. We show the efficacy of our proposed approach for classification and object detection tasks.
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.
Quantization is a promising approach for reducing the inference time and memory footprint of neural networks. However, most existing quantization methods require access to the original training dataset for retraining during quantization. This is often not possible for applications with sensitive or proprietary data, e.g., due to privacy and security concerns. Existing zero-shot quantization methods use different heuristics to address this, but they result in poor performance, especially when quantizing to ultra-low precision. Here, we propose ZeroQ , a novel zero-shot quantization framework to address this. ZeroQ enables mixed-precision quantization without any access to the training or validation data. This is achieved by optimizing for a Distilled Dataset, which is engineered to match the statistics of batch normalization across different layers of the network. ZeroQ supports both uniform and mixed-precision quantization. For the latter, we introduce a novel Pareto frontier based method to automatically determine the mixed-precision bit setting for all layers, with no manual search involved. We extensively test our proposed method on a diverse set of models, including ResNet18/50/152, MobileNetV2, ShuffleNet, SqueezeNext, and InceptionV3 on ImageNet, as well as RetinaNet-ResNet50 on the Microsoft COCO dataset. In particular, we show that ZeroQ can achieve 1.71\% higher accuracy on MobileNetV2, as compared to the recently proposed DFQ method. Importantly, ZeroQ has a very low computational overhead, and it can finish the entire quantization process in less than 30s (0.5\% of one epoch training time of ResNet50 on ImageNet). We have open-sourced the ZeroQ framework\footnote{https://github.com/amirgholami/ZeroQ}.
With the emergence of a spectrum of high-end mobile devices, many applications that formerly required desktop-level computation capability are being transferred to these devices. However, executing the inference of Deep Neural Networks (DNNs) is still challenging considering high computation and storage demands, specifically, if real-time performance with high accuracy is needed. Weight pruning of DNNs is proposed, but existing schemes represent two extremes in the design space: non-structured pruning is fine-grained, accurate, but not hardware friendly; structured pruning is coarse-grained, hardware-efficient, but with higher accuracy loss. In this paper, we introduce a new dimension, fine-grained pruning patterns inside the coarse-grained structures, revealing a previously unknown point in design space. With the higher accuracy enabled by fine-grained pruning patterns, the unique insight is to use the compiler to re-gain and guarantee high hardware efficiency. In other words, our method achieves the best of both worlds, and is desirable across theory/algorithm, compiler, and hardware levels. The proposed PatDNN is an end-to-end framework to efficiently execute DNN on mobile devices with the help of a novel model compression technique (pattern-based pruning based on extended ADMM solution framework) and a set of thorough architecture-aware compiler- and code generation-based optimizations (filter kernel reordering, compressed weight storage, register load redundancy elimination, and parameter auto-tuning). Evaluation results demonstrate that PatDNN outperforms three state-of-the-art end-to-end DNN frameworks, TensorFlow Lite, TVM, and Alibaba Mobile Neural Network with speedup up to 44.5x, 11.4x, and 7.1x, respectively, with no accuracy compromise. Real-time inference of representative large-scale DNNs (e.g., VGG-16, ResNet-50) can be achieved using mobile devices.
In this era of digital information explosion, an abundance of data from numerous modalities is being generated as well as archived everyday. However, most problems associated with training Deep Neural Networks still revolve around lack of data that is rich enough for a given task. Data is required not only for training an initial model, but also for future learning tasks such as Model Compression and Incremental Learning. A diverse dataset may be used for training an initial model, but it may not be feasible to store it throughout the product life cycle due to data privacy issues or memory constraints. We propose to bridge the gap between the abundance of available data and lack of relevant data, for the future learning tasks of a given trained network. We use the available data, that may be an imbalanced subset of the original training dataset, or a related domain dataset, to retrieve representative samples from a trained classifier, using a novel Data-enriching GAN (DeGAN) framework. We demonstrate that data from a related domain can be leveraged to achieve state-of-the-art performance for the tasks of Data-free Knowledge Distillation and Incremental Learning on benchmark datasets. We further demonstrate that our proposed framework can enrich any data, even from unrelated domains, to make it more useful for the future learning tasks of a given network.
Deep Neural Networks (DNNs) have recently been achieving state-of-the-art performance on a variety of computer vision related tasks. However, their computational cost limits their ability to be implemented in embedded systems with restricted resources or strict latency constraints. Model compression has therefore been an active field of research to overcome this issue. Additionally, DNNs typically require massive amounts of labeled data to be trained. This represents a second limitation to their deployment. Domain Adaptation (DA) addresses this issue by allowing knowledge learned on one labeled source distribution to be transferred to a target distribution, possibly unlabeled. In this paper, we investigate on possible improvements of compression methods in DA setting. We focus on a compression method that was previously developed in the context of a single data distribution and show that, with a careful choice of data to use during compression and additional regularization terms directly related to DA objectives, it is possible to improve compression results. We also show that our method outperforms an existing compression method studied in the DA setting by a large margin for high compression rates. Although our work is based on one specific compression method, we also outline some general guidelines for improving compression in DA setting.
Knowledge Distillation (KD) has made remarkable progress in the last few years and become a popular paradigm for model compression and knowledge transfer. However, almost all existing KD algorithms are data-driven, i.e., relying on a large amount of original training data or alternative data, which is usually unavailable in real-world scenarios. In this paper, we devote ourselves to this challenging problem and propose a novel adversarial distillation mechanism to craft a compact student model without any real-world data. We introduce a model discrepancy to quantificationally measure the difference between student and teacher models and construct an optimizable upper bound. In our work, the student and the teacher jointly act the role of the discriminator to reduce this discrepancy, when a generator adversarially produces some "hard samples" to enlarge it. Extensive experiments demonstrate that the proposed data-free method yields comparable performance to existing data-driven methods. More strikingly, our approach can be directly extended to semantic segmentation, which is more complicated than classification, and our approach achieves state-of-the-art results. Code and pretrained models are available at https://github.com/VainF/Data-Free-Adversarial-Distillation.
The success of convolutional neural networks (CNNs) in various applications is accompanied by a significant increase in computation and parameter storage costs. Recent efforts to reduce these overheads involve pruning and compressing the weights of various layers while at the same time aiming to not sacrifice performance. In this paper, we propose a novel criterion for CNN pruning inspired by neural network interpretability: The most relevant units, i.e. weights or filters, are automatically found using their relevance scores obtained from concepts of explainable AI (XAI). By exploring this idea, we connect the lines of interpretability and model compression research. We show that our proposed method can efficiently prune CNN models in transfer-learning setups in which networks pre-trained on large corpora are adapted to specialized tasks. The method is evaluated on a broad range of computer vision datasets. Notably, our novel criterion is not only competitive or better compared to state-of-the-art pruning criteria when successive retraining is performed, but clearly outperforms these previous criteria in the resource-constrained application scenario in which the data of the task to be transferred to is very scarce and one chooses to refrain from fine-tuning. Our method is able to compress the model iteratively while maintaining or even improving accuracy. At the same time, it has a computational cost in the order of gradient computation and is comparatively simple to apply without the need for tuning hyperparameters for pruning.
Counting the number of birds in an open sky setting has been an challenging problem due to the large number of bird flocks and the birds can overlap. Another difficulty is the lack of accurate training samples since the cost of labeling images of bird flocks can be extremely high and each sample picture can contain thousands of birds in a high resolution image. Inspired by recent work on training with synthetic data to perform crowd counting, we design a mechanism to generate synthetic bird dataset with precise bird count and the corresponding density maps. We then train a Unet model on the synthetic dataset to perform density map estimation that produces the count for each input. Our method is able to achieve MSE of approximately 12.4 on real dataset. In order to build a scalable system for fast bird counting under storage and computational constraints, we use model compression techniques and efficient model structures to increase the inference speed and save storage cost. We are able to reduce storage cost from 55MB to less than 5MB for the model with minimum loss of accuracy. This paper describes the pipelines of building an efficient bird counting system.
The computing wall and data movement challenges of deep neural networks (DNNs) have exposed the limitations of conventional CMOS-based DNN accelerators. Furthermore, the deep structure and large model size will make DNNs prohibitive to embedded systems and IoT devices, where low power consumption are required. To address these challenges, spin orbit torque magnetic random-access memory (SOT-MRAM) and SOT-MRAM based Processing-In-Memory (PIM) engines have been used to reduce the power consumption of DNNs since SOT-MRAM has the characteristic of near-zero standby power, high density, none-volatile. However, the drawbacks of SOT-MRAM based PIM engines such as high writing latency and requiring low bit-width data decrease its popularity as a favorable energy efficient DNN accelerator. To mitigate these drawbacks, we propose an ultra energy efficient framework by using model compression techniques including weight pruning and quantization from the software level considering the architecture of SOT-MRAM PIM. And we incorporate the alternating direction method of multipliers (ADMM) into the training phase to further guarantee the solution feasibility and satisfy SOT-MRAM hardware constraints. Thus, the footprint and power consumption of SOT-MRAM PIM can be reduced, while increasing the overall system throughput at the meantime, making our proposed ADMM-based SOT-MRAM PIM more energy efficiency and suitable for embedded systems or IoT devices. Our experimental results show the accuracy and compression rate of our proposed framework is consistently outperforming the reference works, while the efficiency (area \& power) and throughput of SOT-MRAM PIM engine is significantly improved.
Deep Neural Networks have achieved remarkable success relying on the developing high computation capability of GPUs and large-scale datasets with increasing network depth and width in image recognition, object detection and many other applications. However, due to the expensive computation and intensive memory, researchers have concentrated on designing compression methods in recent years. In this paper, we briefly summarize the existing advanced techniques that are useful in model compression at first. After that, we give a detailed description on group lasso regularization and its variants. More importantly, we propose an improving framework of partial regularization based on the relationship between neurons and connections of adjacent layers. It is reasonable and feasible with the help of permutation property of neural network . Experiment results show that partial regularization methods brings improvements such as higher classification accuracy in both training and testing stages on multiple datasets. Since our regularizers contain the computation of less parameters, it shows competitive performances in terms of the total running time of experiments. Finally, we analysed the results and draw a conclusion that the optimal network structure must exist and depend on the input data.
Sequence-level knowledge distillation (SLKD) is a model compression technique that leverages large, accurate teacher models to train smaller, under-parameterized student models. Why does pre-processing MT data with SLKD help us train smaller models? We test the common hypothesis that SLKD addresses a capacity deficiency in students by "simplifying" noisy data points and find it unlikely in our case. Models trained on concatenations of original and "simplified" datasets generalize just as well as baseline SLKD. We then propose an alternative hypothesis under the lens of data augmentation and regularization. We try various augmentation strategies and observe that dropout regularization can become unnecessary. Our methods achieve BLEU gains of 0.7-1.2 on TED Talks.
Besides accuracy, the model size of convolutional neural networks (CNN) models is another important factor considering limited hardware resources in practical applications. For example, employing deep neural networks on mobile systems requires the design of accurate yet fast CNN for low latency in classification and object detection. To fulfill the need, we aim at obtaining CNN models with both high testing accuracy and small size to address resource constraints in many embedded devices. In particular, this paper focuses on proposing a generic reinforcement learning-based model compression approach in a two-stage compression pipeline: pruning and quantization. The first stage of compression, i.e., pruning, is achieved via exploiting deep reinforcement learning (DRL) to co-learn the accuracy and the FLOPs updated after layer-wise channel pruning and element-wise variational pruning via information dropout. The second stage, i.e., quantization, is achieved via a similar DRL approach but focuses on obtaining the optimal bits representation for individual layers. We further conduct experimental results on CIFAR-10 and ImageNet datasets. For the CIFAR-10 dataset, the proposed method can reduce the size of VGGNet by 9x from 20.04MB to 2.2MB with a slight accuracy increase. For the ImageNet dataset, the proposed method can reduce the size of VGG-16 by 33x from 138MB to 4.14MB with no accuracy loss.
Recently, an extensive amount of research has been focused on compressing and accelerating Deep Neural Networks (DNN). So far, high compression rate algorithms require part of the training dataset for a low precision calibration, or a fine-tuning process. However, this requirement is unacceptable when the data is unavailable or contains sensitive information, as in medical and biometric use-cases. We present three methods for generating synthetic samples from trained models. Then, we demonstrate how these samples can be used to calibrate and fine-tune quantized models without using any real data in the process. Our best performing method has a negligible accuracy degradation compared to the original training set. This method, which leverages intrinsic batch normalization layers' statistics of the trained model, can be used to evaluate data similarity. Our approach opens a path towards genuine data-free model compression, alleviating the need for training data during model deployment.
Deep Learning models have become the dominant approach in several areas due to their high performance. Unfortunately, the size and hence computational requirements of operating such models can be considerably high. Therefore, this constitutes a limitation for deployment on memory and battery constrained devices such as mobile phones or embedded systems. To address these limitations, we propose a novel and simple pruning method that compresses neural networks by removing entire filters and neurons according to a global threshold across the network without any pre-calculation of layer sensitivity. The resulting model is compact, non-sparse, with the same accuracy as the non-compressed model, and most importantly requires no special infrastructure for deployment. We prove the viability of our method by producing highly compressed models, namely VGG-16, ResNet-56, and ResNet-110 respectively on CIFAR10 without losing any performance compared to the baseline, as well as ResNet-34 and ResNet-50 on ImageNet without a significant loss of accuracy. We also provide a well-retrained 30% compressed ResNet-50 that slightly surpasses the base model accuracy. Additionally, compressing more than 56% and 97% of AlexNet and LeNet-5 respectively. Interestingly, the resulted models' pruning patterns are highly similar to the other methods using layer sensitivity pre-calculation step. Our method does not only exhibit good performance but what is more also easy to implement.
While state-of-the-art NLP explainability (XAI) methods focus on explaining per-sample decisions in supervised end or probing tasks, this is insufficient to explain and quantify model knowledge transfer during (un-)supervised training. Thus, for TX-Ray, we modify the established computer vision explainability principle of 'visualizing preferred inputs of neurons' to make it usable transfer analysis and NLP. This allows one to analyze, track and quantify how self- or supervised NLP models first build knowledge abstractions in pretraining (1), and then transfer these abstractions to a new domain (2), or adapt them during supervised fine-tuning (3). TX-Ray expresses neurons as feature preference distributions to quantify fine-grained knowledge transfer or adaptation and guide human analysis. We find that, similar to Lottery Ticket based pruning, TX-Ray based pruning can improve test set generalization and that it can reveal how early stages of self-supervision automatically learn linguistic abstractions like parts-of-speech.
We propose an efficient distributed online learning protocol for low-latency real-time services. It extends a previously presented protocol to kernelized online learners that represent their models by a support vector expansion. While such learners often achieve higher predictive performance than their linear counterparts, communicating the support vector expansions becomes inefficient for large numbers of support vectors. The proposed extension allows for a larger class of online learning algorithms---including those alleviating the problem above through model compression. In addition, we characterize the quality of the proposed protocol by introducing a novel criterion that requires the communication to be bounded by the loss suffered.
Deploying trained convolutional neural networks (CNNs) to mobile devices is a challenging task because of the simultaneous requirements of the deployed model to be fast, lightweight and accurate. Designing and training a CNN architecture that does well on all three metrics is highly non-trivial and can be very time-consuming if done by hand. One way to solve this problem is to compress the trained CNN models before deploying to mobile devices. This work asks and answers three questions on compressing CNN models automatically: a) How to control the trade-off between speed, memory and accuracy during model compression? b) In practice, a deployed model may not see all classes and/or may not need to produce all class labels. Can this fact be used to improve the trade-off? c) How to scale the compression algorithm to execute within a reasonable amount of time for many deployments? The paper demonstrates that a model compression algorithm utilizing reinforcement learning with architecture search and knowledge distillation can answer these questions in the affirmative. Experimental results are provided for current state-of-the-art CNN model families for image feature extraction like VGG and ResNet with CIFAR datasets.
Despite the success of deep neural networks (DNNs), state-of-the-art models are too large to deploy on low-resource devices or common server configurations in which multiple models are held in memory. Model compression methods address this limitation by reducing the memory footprint, latency, or energy consumption of a model with minimal impact on accuracy. We focus on the task of reducing the number of learnable variables in the model. In this work we combine ideas from weight hashing and dimensionality reductions resulting in a simple and powerful structured multi-hashing method based on matrix products that allows direct control of model size of any deep network and is trained end-to-end. We demonstrate the strength of our approach by compressing models from the ResNet, EfficientNet, and MobileNet architecture families. Our method allows us to drastically decrease the number of variables while maintaining high accuracy. For instance, by applying our approach to EfficentNet-B4 (16M parameters) we reduce it to to the size of B0 (5M parameters), while gaining over 3% in accuracy over B0 baseline. On the commonly used benchmark CIFAR10 we reduce the ResNet32 model by 75% with no loss in quality, and are able to do a 10x compression while still achieving above 90% accuracy.