PyTorchKR Papers

최신 논문

총 5,575편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:1909.01861

Rethinking the Number of Channels for the Convolutional Neural Network

Latest algorithms for automatic neural architecture search perform remarkable but few of them can effectively design the number of channels for convolutional neural networks and consume less computational efforts. In this paper, we propose a method for efficient automatic architecture search which is special to the widths of networks instead of the connections of neural architecture. Our method, functionally incremental search based on function-preserving, will explore the number of channels rapidly while controlling the number of parameters of the target network. On CIFAR-10 and CIFAR-100 classification, our method using minimal computational resources (0.4~1.3 GPU-days) can discover more efficient rules of the widths of networks to improve the accuracy by about 0.5% on CIFAR-10 and a~2.33% on CIFAR-100 with fewer number of parameters. In particular, our method is suitable for exploring the number of channels of almost any convolutional neural network rapidly.

arXiv:1909.01786

GPU-based parallelism for ASP-solving

Answer Set Programming (ASP) has become, the paradigm of choice in the field of logic programming and non-monotonic reasoning. Thanks to the availability of efficient solvers, ASP has been successfully employed in a large number of application domains. The term GPU-computing indicates a recent programming paradigm aimed at enabling the use of modern parallel Graphical Processing Units (GPUs) for general purpose computing. In this paper we describe an approach to ASP-solving that exploits GPU parallelism. The design of a GPU-based solver poses various challenges due to the peculiarities of GPUs' software and hardware architectures and to the intrinsic nature of the satisfiability problem.

arXiv:1909.01754

An Efficient and Layout-Independent Automatic License Plate Recognition System Based on the YOLO detector

This paper presents an efficient and layout-independent Automatic License Plate Recognition (ALPR) system based on the state-of-the-art YOLO object detector that contains a unified approach for license plate (LP) detection and layout classification to improve the recognition results using post-processing rules. The system is conceived by evaluating and optimizing different models, aiming at achieving the best speed/accuracy trade-off at each stage. The networks are trained using images from several datasets, with the addition of various data augmentation techniques, so that they are robust under different conditions. The proposed system achieved an average end-to-end recognition rate of 96.9% across eight public datasets (from five different regions) used in the experiments, outperforming both previous works and commercial systems in the ChineseLP, OpenALPR-EU, SSIG-SegPlate and UFPR-ALPR datasets. In the other datasets, the proposed approach achieved competitive results to those attained by the baselines. Our system also achieved impressive frames per second (FPS) rates on a high-end GPU, being able to perform in real time even when there are four vehicles in the scene. An additional contribution is that we manually labeled 38,351 bounding boxes on 6,239 images from public datasets and made the annotations publicly available to the research community.

arXiv:1909.00548

Resource Optimized Neural Architecture Search for 3D Medical Image Segmentation

Neural Architecture Search (NAS), a framework which automates the task of designing neural networks, has recently been actively studied in the field of deep learning. However, there are only a few NAS methods suitable for 3D medical image segmentation. Medical 3D images are generally very large; thus it is difficult to apply previous NAS methods due to their GPU computational burden and long training time. We propose the resource-optimized neural architecture search method which can be applied to 3D medical segmentation tasks in a short training time (1.39 days for 1GB dataset) using a small amount of computation power (one RTX 2080Ti, 10.8GB GPU memory). Excellent performance can also be achieved without retraining(fine-tuning) which is essential in most NAS methods. These advantages can be achieved by using a reinforcement learning-based controller with parameter sharing and focusing on the optimal search space configuration of macro search rather than micro search. Our experiments demonstrate that the proposed NAS method outperforms manually designed networks with state-of-the-art performance in 3D medical image segmentation.

arXiv:1909.00562

Hybrid Data-Model Parallel Training for Sequence-to-Sequence Recurrent Neural Network Machine Translation

Reduction of training time is an important issue in many tasks like patent translation involving neural networks. Data parallelism and model parallelism are two common approaches for reducing training time using multiple graphics processing units (GPUs) on one machine. In this paper, we propose a hybrid data-model parallel approach for sequence-to-sequence (Seq2Seq) recurrent neural network (RNN) machine translation. We apply a model parallel approach to the RNN encoder-decoder part of the Seq2Seq model and a data parallel approach to the attention-softmax part of the model. We achieved a speed-up of 4.13 to 4.20 times when using 4 GPUs compared with the training speed when using 1 GPU without affecting machine translation accuracy as measured in terms of BLEU scores.

arXiv:1908.09000

Foveated image processing for faster object detection and recognition in embedded systems using deep convolutional neural networks

Object detection and recognition algorithms using deep convolutional neural networks (CNNs) tend to be computationally intensive to implement. This presents a particular challenge for embedded systems, such as mobile robots, where the computational resources tend to be far less than for workstations. As an alternative to standard, uniformly sampled images, we propose the use of foveated image sampling here to reduce the size of images, which are faster to process in a CNN due to the reduced number of convolution operations. We evaluate object detection and recognition on the Microsoft COCO database, using foveated image sampling at different image sizes, ranging from 416x416 to 96x96 pixels, on an embedded GPU -- an NVIDIA Jetson TX2 with 256 CUDA cores. The results show that it is possible to achieve a 4x speed-up in frame rates, from 3.59 FPS to 15.24 FPS, using 416x416 and 128x128 pixel images respectively. For foveated sampling, this image size reduction led to just a small decrease in recall performance in the foveal region, to 92.0% of the baseline performance with full-sized images, compared to a significant decrease to 50.1% of baseline recall performance in uniformly sampled images, demonstrating the advantage of foveated sampling.

arXiv:1908.11848

Dynamic Stale Synchronous Parallel Distributed Training for Deep Learning

Deep learning is a popular machine learning technique and has been applied to many real-world problems. However, training a deep neural network is very time-consuming, especially on big data. It has become difficult for a single machine to train a large model over large datasets. A popular solution is to distribute and parallelize the training process across multiple machines using the parameter server framework. In this paper, we present a distributed paradigm on the parameter server framework called Dynamic Stale Synchronous Parallel (DSSP) which improves the state-of-the-art Stale Synchronous Parallel (SSP) paradigm by dynamically determining the staleness threshold at the run time. Conventionally to run distributed training in SSP, the user needs to specify a particular staleness threshold as a hyper-parameter. However, a user does not usually know how to set the threshold and thus often finds a threshold value through trial and error, which is time-consuming. Based on workers' recent processing time, our approach DSSP adaptively adjusts the threshold per iteration at running time to reduce the waiting time of faster workers for synchronization of the globally shared parameters, and consequently increases the frequency of parameters updates (increases iteration throughput), which speedups the convergence rate. We compare DSSP with other paradigms such as Bulk Synchronous Parallel (BSP), Asynchronous Parallel (ASP), and SSP by running deep neural networks (DNN) models over GPU clusters in both homogeneous and heterogeneous environments. The results show that in a heterogeneous environment where the cluster consists of mixed models of GPUs, DSSP converges to a higher accuracy much earlier than SSP and BSP and performs similarly to ASP. In a homogeneous distributed cluster, DSSP has more stable and slightly better performance than SSP and ASP, and converges much faster than BSP.

arXiv:1908.11809

High Performance Scalable FPGA Accelerator for Deep Neural Networks

Low-precision is the first order knob for achieving higher Artificial Intelligence Operations (AI-TOPS). However the algorithmic space for sub-8-bit precision compute is diverse, with disruptive changes happening frequently, making FPGAs a natural choice for Deep Neural Network inference, In this work we present an FPGA-based accelerator for CNN inference acceleration. We use {\it INT-8-2} compute (with {\it 8 bit} activation and {2 bit} weights) which is recently showing promise in the literature, and which no known ASIC, CPU or GPU natively supports today. Using a novel Adaptive Logic Module (ALM) based design, as a departure from traditional DSP based designs, we are able to achieve high performance measurement of 5 AI-TOPS for {\it Arria10} and project a performance of 76 AI-TOPS at 0.7 TOPS/W for {\it Stratix10}. This exceeds known CPU, GPU performance and comes close to best known ASIC (TPU) numbers, while retaining the versatility of the FPGA platform for other applications.

arXiv:1908.11450

Neural Network Inference on Mobile SoCs

The ever-increasing demand from mobile Machine Learning (ML) applications calls for evermore powerful on-chip computing resources. Mobile devices are empowered with heterogeneous multi-processor Systems-on-Chips (SoCs) to process ML workloads such as Convolutional Neural Network (CNN) inference. Mobile SoCs house several different types of ML capable components on-die, such as CPU, GPU, and accelerators. These different components are capable of independently performing inference but with very different power-performance characteristics. In this article, we provide a quantitative evaluation of the inference capabilities of the different components on mobile SoCs. We also present insights behind their respective power-performance behavior. Finally, we explore the performance limit of the mobile SoCs by synergistically engaging all the components concurrently. We observe that a mobile SoC provides up to 2x improvement with parallel inference when all its components are engaged, as opposed to engaging only one component.

arXiv:1908.11656

LU-Net: An Efficient Network for 3D LiDAR Point Cloud Semantic Segmentation Based on End-to-End-Learned 3D Features and U-Net

We propose LU-Net -- for LiDAR U-Net, a new method for the semantic segmentation of a 3D LiDAR point cloud. Instead of applying some global 3D segmentation method such as PointNet, we propose an end-to-end architecture for LiDAR point cloud semantic segmentation that efficiently solves the problem as an image processing problem. We first extract high-level 3D features for each point given its 3D neighbors. Then, these features are projected into a 2D multichannel range-image by considering the topology of the sensor. Thanks to these learned features and this projection, we can finally perform the segmentation using a simple U-Net segmentation network, which performs very well while being very efficient. In this way, we can exploit both the 3D nature of the data and the specificity of the LiDAR sensor. This approach outperforms the state-of-the-art by a large margin on the KITTI dataset, as our experiments show. Moreover, this approach operates at 24fps on a single GPU. This is above the acquisition rate of common LiDAR sensors which makes it suitable for real-time applications.

arXiv:1908.10714

Automated Architecture Design for Deep Neural Networks

Machine learning has made tremendous progress in recent years and received large amounts of public attention. Though we are still far from designing a full artificially intelligent agent, machine learning has brought us many applications in which computers solve human learning tasks remarkably well. Much of this progress comes from a recent trend within machine learning, called deep learning. Deep learning models are responsible for many state-of-the-art applications of machine learning. Despite their success, deep learning models are hard to train, very difficult to understand, and often times so complex that training is only possible on very large GPU clusters. Lots of work has been done on enabling neural networks to learn efficiently. However, the design and architecture of such neural networks is often done manually through trial and error and expert knowledge. This thesis inspects different approaches, existing and novel, to automate the design of deep feedforward neural networks in an attempt to create less complex models with good performance that take away the burden of deciding on an architecture and make it more efficient to design and train such deep networks.

arXiv:1908.09207

Demystifying the MLPerf Benchmark Suite

MLPerf, an emerging machine learning benchmark suite strives to cover a broad range of applications of machine learning. We present a study on its characteristics and how the MLPerf benchmarks differ from some of the previous deep learning benchmarks like DAWNBench and DeepBench. We find that application benchmarks such as MLPerf (although rich in kernels) exhibit different features compared to kernel benchmarks such as DeepBench. MLPerf benchmark suite contains a diverse set of models which allows unveiling various bottlenecks in the system. Based on our findings, dedicated low latency interconnect between GPUs in multi-GPU systems is required for optimal distributed deep learning training. We also observe variation in scaling efficiency across the MLPerf models. The variation exhibited by the different models highlight the importance of smart scheduling strategies for multi-GPU training. Another observation is that CPU utilization increases with increase in number of GPUs used for training. Corroborating prior work we also observe and quantify improvements possible by compiler optimizations, mixed-precision training and use of Tensor Cores.

arXiv:1908.09999공식 코드 · Papers with Code

A2J: Anchor-to-Joint Regression Network for 3D Articulated Pose Estimation from a Single Depth Image

For 3D hand and body pose estimation task in depth image, a novel anchor-based approach termed Anchor-to-Joint regression network (A2J) with the end-to-end learning ability is proposed. Within A2J, anchor points able to capture global-local spatial context information are densely set on depth image as local regressors for the joints. They contribute to predict the positions of the joints in ensemble way to enhance generalization ability. The proposed 3D articulated pose estimation paradigm is different from the state-of-the-art encoder-decoder based FCN, 3D CNN and point-set based manners. To discover informative anchor points towards certain joint, anchor proposal procedure is also proposed for A2J. Meanwhile 2D CNN (i.e., ResNet-50) is used as backbone network to drive A2J, without using time-consuming 3D convolutional or deconvolutional layers. The experiments on 3 hand datasets and 2 body datasets verify A2J's superiority. Meanwhile, A2J is of high running speed around 100 FPS on single NVIDIA 1080Ti GPU.

arXiv:1908.09791공식 코드 · Papers with Code

Once-for-All: Train One Network and Specialize it for Efficient Deployment

We address the challenging problem of efficient inference across many devices and resource constraints, especially on edge devices. Conventional approaches either manually design or use neural architecture search (NAS) to find a specialized neural network and train it from scratch for each case, which is computationally prohibitive (causing $CO_2$ emission as much as 5 cars' lifetime) thus unscalable. In this work, we propose to train a once-for-all (OFA) network that supports diverse architectural settings by decoupling training and search, to reduce the cost. We can quickly get a specialized sub-network by selecting from the OFA network without additional training. To efficiently train OFA networks, we also propose a novel progressive shrinking algorithm, a generalized pruning method that reduces the model size across many more dimensions than pruning (depth, width, kernel size, and resolution). It can obtain a surprisingly large number of sub-networks ($> 10^{19}$) that can fit different hardware platforms and latency constraints while maintaining the same level of accuracy as training independently. On diverse edge devices, OFA consistently outperforms state-of-the-art (SOTA) NAS methods (up to 4.0% ImageNet top1 accuracy improvement over MobileNetV3, or same accuracy but 1.5x faster than MobileNetV3, 2.6x faster than EfficientNet w.r.t measured latency) while reducing many orders of magnitude GPU hours and $CO_2$ emission. In particular, OFA achieves a new SOTA 80.0% ImageNet top-1 accuracy under the mobile setting ($<$600M MACs). OFA is the winning solution for the 3rd Low Power Computer Vision Challenge (LPCVC), DSP classification track and the 4th LPCVC, both classification track and detection track. Code and 50 pre-trained models (for many devices & many latency constraints) are released at https://github.com/mit-han-lab/once-for-all.

arXiv:1908.09550

Customizable Architecture Search for Semantic Segmentation

In this paper, we propose a Customizable Architecture Search (CAS) approach to automatically generate a network architecture for semantic image segmentation. The generated network consists of a sequence of stacked computation cells. A computation cell is represented as a directed acyclic graph, in which each node is a hidden representation (i.e., feature map) and each edge is associated with an operation (e.g., convolution and pooling), which transforms data to a new layer. During the training, the CAS algorithm explores the search space for an optimized computation cell to build a network. The cells of the same type share one architecture but with different weights. In real applications, however, an optimization may need to be conducted under some constraints such as GPU time and model size. To this end, a cost corresponding to the constraint will be assigned to each operation. When an operation is selected during the search, its associated cost will be added to the objective. As a result, our CAS is able to search an optimized architecture with customized constraints. The approach has been thoroughly evaluated on Cityscapes and CamVid datasets, and demonstrates superior performance over several state-of-the-art techniques. More remarkably, our CAS achieves 72.3% mIoU on the Cityscapes dataset with speed of 108 FPS on an Nvidia TitanXp GPU.

arXiv:1908.08024

Mapping of Local and Global Synapses on Spiking Neuromorphic Hardware

Spiking Neural Networks (SNNs) are widely deployed to solve complex pattern recognition, function approximation and image classification tasks. With the growing size and complexity of these networks, hardware implementation becomes challenging because scaling up the size of a single array (crossbar) of fully connected neurons is no longer feasible due to strict energy budget. Modern neromorphic hardware integrates small-sized crossbars with time-multiplexed interconnects. Partitioning SNNs becomes essential in order to map them on neuromorphic hardware with the major aim to reduce the global communication latency and energy overhead. To achieve this goal, we propose our instantiation of particle swarm optimization, which partitions SNNs into local synapses (mapped on crossbars) and global synapses (mapped on time-multiplexed interconnects), with the objective of reducing spike communication on the interconnect. This improves latency, power consumption as well as application performance by reducing inter-spike interval distortion and spike disorders. Our framework is implemented in Python, interfacing CARLsim, a GPU-accelerated application-level spiking neural network simulator with an extended version of Noxim, for simulating time-multiplexed interconnects. Experiments are conducted with realistic and synthetic SNN-based applications with different computation models, topologies and spike coding schemes. Using power numbers from in-house neuromorphic chips, we demonstrate significant reductions in energy consumption and spike latency over PACMAN, the widely-used partitioning technique for SNNs on SpiNNaker.

arXiv:1908.07847

CUDA optimized Neural Network predicts blood glucose control from quantified joint mobility and anthropometrics

Neural network training entails heavy computation with obvious bottlenecks. The Compute Unified Device Architecture (CUDA) programming model allows us to accelerate computation by passing the processing workload from the CPU to the graphics processing unit (GPU). In this paper, we leveraged the power of Nvidia GPUs to parallelize all of the computation involved in training, to accelerate a backpropagation feed-forward neural network with one hidden layer using CUDA and C++. This optimized neural network was tasked with predicting the level of glycated hemoglobin (HbA1c) from non-invasive markers. The rate of increase in the prevalence of Diabetes Mellitus has resulted in an urgent need for early detection and accurate diagnosis. However, due to the invasiveness and limitations of conventional tests, alternate means are being considered. Limited Joint Mobility (LJM) has been reported as an indicator for poor glycemic control. LJM of the fingers is quantified and its link to HbA1c is investigated along with other potential non-invasive markers of HbA1c. We collected readings of 33 potential markers from 120 participants at a clinic in south Trinidad. Our neural network achieved 95.65% accuracy on the training and 86.67% accuracy on the testing set for male participants and 97.73% and 66.67% accuracy on the training and testing sets for female participants. Using 960 CUDA cores from a Nvidia GeForce GTX 660, our parallelized neural network was trained 50 times faster on both subsets, than its corresponding CPU implementation on an Intel Core (TM) i7-3630QM 2.40 GHz CPU.

arXiv:1908.07678공식 코드 · Papers with Code

Asymmetric Non-local Neural Networks for Semantic Segmentation

The non-local module works as a particularly useful technique for semantic segmentation while criticized for its prohibitive computation and GPU memory occupation. In this paper, we present Asymmetric Non-local Neural Network to semantic segmentation, which has two prominent components: Asymmetric Pyramid Non-local Block (APNB) and Asymmetric Fusion Non-local Block (AFNB). APNB leverages a pyramid sampling module into the non-local block to largely reduce the computation and memory consumption without sacrificing the performance. AFNB is adapted from APNB to fuse the features of different levels under a sufficient consideration of long range dependencies and thus considerably improves the performance. Extensive experiments on semantic segmentation benchmarks demonstrate the effectiveness and efficiency of our work. In particular, we report the state-of-the-art performance of 81.3 mIoU on the Cityscapes test set. For a 256x128 input, APNB is around 6 times faster than a non-local block on GPU while 28 times smaller in GPU running memory occupation. Code is available at: https://github.com/MendelXu/ANN.git.

arXiv:1908.06972

PrivFT: Private and Fast Text Classification with Homomorphic Encryption

The need for privacy-preserving analytics is higher than ever due to the severity of privacy risks and to comply with new privacy regulations leading to an amplified interest in privacy-preserving techniques that try to balance between privacy and utility. In this work, we present an efficient method for Text Classification while preserving the privacy of the content using Fully Homomorphic Encryption (FHE). Our system (named \textbf{Priv}ate \textbf{F}ast \textbf{T}ext (PrivFT)) performs two tasks: 1) making inference of encrypted user inputs using a plaintext model and 2) training an effective model using an encrypted dataset. For inference, we train a supervised model and outline a system for homomorphic inference on encrypted user inputs with zero loss to prediction accuracy. In the second part, we show how to train a model using fully encrypted data to generate an encrypted model. We provide a GPU implementation of the Cheon-Kim-Kim-Song (CKKS) FHE scheme and compare it with existing CPU implementations to achieve 1 to 2 orders of magnitude speedup at various parameter settings. We implement PrivFT in GPUs to achieve a run time per inference of less than 0.66 seconds. Training on a relatively large encrypted dataset is more computationally intensive requiring 5.04 days.

arXiv:1908.05599

Deep Slice Interpolation via Marginal Super-Resolution, Fusion and Refinement

We propose a marginal super-resolution (MSR) approach based on 2D convolutional neural networks (CNNs) for interpolating an anisotropic brain magnetic resonance scan along the highly under-sampled direction, which is assumed to axial without loss of generality. Previous methods for slice interpolation only consider data from pairs of adjacent 2D slices. The possibility of fusing information from the direction orthogonal to the 2D slices remains unexplored. Our approach performs MSR in both sagittal and coronal directions, which provides an initial estimate for slice interpolation. The interpolated slices are then fused and refined in the axial direction for improved consistency. Since MSR consists of only 2D operations, it is more feasible in terms of GPU memory consumption and requires fewer training samples compared to 3D CNNs. Our experiments demonstrate that the proposed method outperforms traditional linear interpolation and baseline 2D/3D CNN-based approaches. We conclude by showcasing the method's practical utility in estimating brain volumes from under-sampled brain MR scans through semantic segmentation.