PyTorchKR Papers

최신 논문

총 5,576편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2101.00265

VisualSparta: An Embarrassingly Simple Approach to Large-scale Text-to-Image Search with Weighted Bag-of-words

Text-to-image retrieval is an essential task in cross-modal information retrieval, i.e., retrieving relevant images from a large and unlabelled dataset given textual queries. In this paper, we propose VisualSparta, a novel (Visual-text Sparse Transformer Matching) model that shows significant improvement in terms of both accuracy and efficiency. VisualSparta is capable of outperforming previous state-of-the-art scalable methods in MSCOCO and Flickr30K. We also show that it achieves substantial retrieving speed advantages, i.e., for a 1 million image index, VisualSparta using CPU gets ~391X speedup compared to CPU vector search and ~5.4X speedup compared to vector search with GPU acceleration. Experiments show that this speed advantage even gets bigger for larger datasets because VisualSparta can be efficiently implemented as an inverted index. To the best of our knowledge, VisualSparta is the first transformer-based text-to-image retrieval model that can achieve real-time searching for large-scale datasets, with significant accuracy improvement compared to previous state-of-the-art methods.

arXiv:2101.00793

A Framework for Fast Scalable BNN Inference using Googlenet and Transfer Learning

Efficient and accurate object detection in video and image analysis is one of the major beneficiaries of the advancement in computer vision systems with the help of deep learning. With the aid of deep learning, more powerful tools evolved, which are capable to learn high-level and deeper features and thus can overcome the existing problems in traditional architectures of object detection algorithms. The work in this thesis aims to achieve high accuracy in object detection with good real-time performance. In the area of computer vision, a lot of research is going into the area of detection and processing of visual information, by improving the existing algorithms. The binarized neural network has shown high performance in various vision tasks such as image classification, object detection, and semantic segmentation. The Modified National Institute of Standards and Technology database (MNIST), Canadian Institute for Advanced Research (CIFAR), and Street View House Numbers (SVHN) datasets are used which is implemented using a pre-trained convolutional neural network (CNN) that is 22 layers deep. Supervised learning is used in the work, which classifies the particular dataset with the proper structure of the model. In still images, to improve accuracy, Googlenet is used. The final layer of the Googlenet is replaced with the transfer learning to improve the accuracy of the Googlenet. At the same time, the accuracy in moving images can be maintained by transfer learning techniques. Hardware is the main backbone for any model to obtain faster results with a large number of datasets. Here, Nvidia Jetson Nano is used which is a graphics processing unit (GPU), that can handle a large number of computations in the process of object detection. Results show that the accuracy of objects detected by the transfer learning method is more when compared to the existing methods.

arXiv:2101.00732

Generalized Latency Performance Estimation for Once-For-All Neural Architecture Search

Neural Architecture Search (NAS) has enabled the possibility of automated machine learning by streamlining the manual development of deep neural network architectures defining a search space, search strategy, and performance estimation strategy. To solve the need for multi-platform deployment of Convolutional Neural Network (CNN) models, Once-For-All (OFA) proposed to decouple Training and Search to deliver a one-shot model of sub-networks that are constrained to various accuracy-latency tradeoffs. We find that the performance estimation strategy for OFA's search severely lacks generalizability of different hardware deployment platforms due to single hardware latency lookup tables that require significant amount of time and manual effort to build beforehand. In this work, we demonstrate the framework for building latency predictors for neural network architectures to address the need for heterogeneous hardware support and reduce the overhead of lookup tables altogether. We introduce two generalizability strategies which include fine-tuning using a base model trained on a specific hardware and NAS search space, and GPU-generalization which trains a model on GPU hardware parameters such as Number of Cores, RAM Size, and Memory Bandwidth. With this, we provide a family of latency prediction models that achieve over 50% lower RMSE loss as compared to with ProxylessNAS. We also show that the use of these latency predictors match the NAS performance of the lookup table baseline approach if not exceeding it in certain cases.

arXiv:2101.00745공식 코드 · Papers with Code

DSXplore: Optimizing Convolutional Neural Networks via Sliding-Channel Convolutions

As the key advancement of the convolutional neural networks (CNNs), depthwise separable convolutions (DSCs) are becoming one of the most popular techniques to reduce the computations and parameters size of CNNs meanwhile maintaining the model accuracy. It also brings profound impact to improve the applicability of the compute- and memory-intensive CNNs to a broad range of applications, such as mobile devices, which are generally short of computation power and memory. However, previous research in DSCs are largely focusing on compositing the limited existing DSC designs, thus, missing the opportunities to explore more potential designs that can achieve better accuracy and higher computation/parameter reduction. Besides, the off-the-shelf convolution implementations offer limited computing schemes, therefore, lacking support for DSCs with different convolution patterns. To this end, we introduce, DSXplore, the first optimized design for exploring DSCs on CNNs. Specifically, at the algorithm level, DSXplore incorporates a novel factorized kernel -- sliding-channel convolution (SCC), featured with input-channel overlapping to balance the accuracy performance and the reduction of computation and memory cost. SCC also offers enormous space for design exploration by introducing adjustable kernel parameters. Further, at the implementation level, we carry out an optimized GPU-implementation tailored for SCC by leveraging several key techniques, such as the input-centric backward design and the channel-cyclic optimization. Intensive experiments on different datasets across mainstream CNNs show the advantages of DSXplore in balancing accuracy and computation/parameter reduction over the standard convolution and the existing DSCs.

arXiv:2012.15667

I/O Lower Bounds for Auto-tuning of Convolutions in CNNs

Convolution is the most time-consuming part in the computation of convolutional neural networks (CNNs), which have achieved great successes in numerous applications. Due to the complex data dependency and the increase in the amount of model samples, the convolution suffers from high overhead on data movement (i.e., memory access). This work provides comprehensive analysis and methodologies to minimize the communication for the convolution in CNNs. With an in-depth analysis of the recent I/O complexity theory under the red-blue game model, we develop a general I/O lower bound theory for a composite algorithm which consists of several different sub-computations. Based on the proposed theory, we establish the data movement lower bound results of two representative convolution algorithms in CNNs, namely the direct convolution and Winograd algorithm. Next, derived from I/O lower bound results, we design the near I/O-optimal dataflow strategies for the two main convolution algorithms by fully exploiting the data reuse. Furthermore, in order to push the envelope of performance of the near I/O-optimal dataflow strategies further, an aggressive design of auto-tuning based on I/O lower bounds, is proposed to search an optimal parameter configuration for the direct convolution and Winograd algorithm on GPU, such as the number of threads and the size of shared memory used in each thread block. Finally, experiment evaluation results on the direct convolution and Winograd algorithm show that our dataflow strategies with the auto-tuning approach can achieve about 3.32x performance speedup on average over cuDNN. In addition, compared with TVM, which represents the state-of-the-art technique for auto-tuning, not only our auto-tuning method based on I/O lower bounds can find the optimal parameter configuration faster, but also our solution has higher performance than the optimal solution provided by TVM.

arXiv:2012.15712공식 코드 · Papers with Code

Voxel R-CNN: Towards High Performance Voxel-based 3D Object Detection

Recent advances on 3D object detection heavily rely on how the 3D data are represented, \emph{i.e.}, voxel-based or point-based representation. Many existing high performance 3D detectors are point-based because this structure can better retain precise point positions. Nevertheless, point-level features lead to high computation overheads due to unordered storage. In contrast, the voxel-based structure is better suited for feature extraction but often yields lower accuracy because the input data are divided into grids. In this paper, we take a slightly different viewpoint -- we find that precise positioning of raw points is not essential for high performance 3D object detection and that the coarse voxel granularity can also offer sufficient detection accuracy. Bearing this view in mind, we devise a simple but effective voxel-based framework, named Voxel R-CNN. By taking full advantage of voxel features in a two stage approach, our method achieves comparable detection accuracy with state-of-the-art point-based models, but at a fraction of the computation cost. Voxel R-CNN consists of a 3D backbone network, a 2D bird-eye-view (BEV) Region Proposal Network and a detect head. A voxel RoI pooling is devised to extract RoI features directly from voxel features for further refinement. Extensive experiments are conducted on the widely used KITTI Dataset and the more recent Waymo Open Dataset. Our results show that compared to existing voxel-based methods, Voxel R-CNN delivers a higher detection accuracy while maintaining a real-time frame processing rate, \emph{i.e}., at a speed of 25 FPS on an NVIDIA RTX 2080 Ti GPU. The code is available at \url{https://github.com/djiajunustc/Voxel-R-CNN}.

arXiv:2012.14539

TensorX: Extensible API for Neural Network Model Design and Deployment

TensorX is a Python library for prototyping, design, and deployment of complex neural network models in TensorFlow. A special emphasis is put on ease of use, performance, and API consistency. It aims to make available high-level components like neural network layers that are, in effect, stateful functions, easy to compose and reuse. Its architecture allows for the expression of patterns commonly found when building neural network models either on research or industrial settings. Incorporating ideas from several other deep learning libraries, it makes it easy to use components commonly found in state-of-the-art models. The library design mixes functional dataflow computation graphs with object-oriented neural network building blocks. TensorX combines the dynamic nature of Python with the high-performance GPU-enabled operations of TensorFlow. This library has minimal core dependencies (TensorFlow and NumPy) and is distributed under Apache License 2.0 licence, encouraging its use in both an academic and commercial settings. Full documentation, source code, and binaries can be found in https://tensorx.org/.

arXiv:2012.15671공식 코드 · Papers with Code

Vocabulary Learning via Optimal Transport for Neural Machine Translation

The choice of token vocabulary affects the performance of machine translation. This paper aims to figure out what is a good vocabulary and whether one can find the optimal vocabulary without trial training. To answer these questions, we first provide an alternative understanding of the role of vocabulary from the perspective of information theory. Motivated by this, we formulate the quest of vocabularization -- finding the best token dictionary with a proper size -- as an optimal transport (OT) problem. We propose VOLT, a simple and efficient solution without trial training. Empirical results show that VOLT outperforms widely-used vocabularies in diverse scenarios, including WMT-14 English-German and TED's 52 translation directions. For example, VOLT achieves almost 70% vocabulary size reduction and 0.5 BLEU gain on English-German translation. Also, compared to BPE-search, VOLT reduces the search time from 384 GPU hours to 30 GPU hours on English-German translation. Codes are available at https://github.com/Jingjing-NLP/VOLT .

arXiv:2012.15024

Adaptive Graph Diffusion Networks

Graph Neural Networks (GNNs) have received much attention in the graph deep learning domain. However, recent research empirically and theoretically shows that deep GNNs suffer from over-fitting and over-smoothing problems. The usual solutions either cannot solve extensive runtime of deep GNNs or restrict graph convolution in the same feature space. We propose the Adaptive Graph Diffusion Networks (AGDNs) which perform multi-layer generalized graph diffusion in different feature spaces with moderate complexity and runtime. Standard graph diffusion methods combine large and dense powers of the transition matrix with predefined weighting coefficients. Instead, AGDNs combine smaller multi-hop node representations with learnable and generalized weighting coefficients. We propose two scalable mechanisms of weighting coefficients to capture multi-hop information: Hop-wise Attention (HA) and Hop-wise Convolution (HC). We evaluate AGDNs on diverse, challenging Open Graph Benchmark (OGB) datasets with semi-supervised node classification and link prediction tasks. Until the date of submission (Aug 26, 2022), AGDNs achieve top-1 performance on the ogbn-arxiv, ogbn-proteins and ogbl-ddi datasets and top-3 performance on the ogbl-citation2 dataset. On the similar Tesla V100 GPU cards, AGDNs outperform Reversible GNNs (RevGNNs) with 13% complexity and 1% training runtime of RevGNNs on the ogbn-proteins dataset. AGDNs also achieve comparable performance to SEAL with 36% training and 0.2% inference runtime of SEAL on the ogbl-citation2 dataset.

arXiv:2012.15843

A Tale of Two Efficient and Informative Negative Sampling Distributions

Softmax classifiers with a very large number of classes naturally occur in many applications such as natural language processing and information retrieval. The calculation of full softmax is costly from the computational and energy perspective. There have been various sampling approaches to overcome this challenge, popularly known as negative sampling (NS). Ideally, NS should sample negative classes from a distribution that is dependent on the input data, the current parameters, and the correct positive class. Unfortunately, due to the dynamically updated parameters and data samples, there is no sampling scheme that is provably adaptive and samples the negative classes efficiently. Therefore, alternative heuristics like random sampling, static frequency-based sampling, or learning-based biased sampling, which primarily trade either the sampling cost or the adaptivity of samples per iteration are adopted. In this paper, we show two classes of distributions where the sampling scheme is truly adaptive and provably generates negative samples in near-constant time. Our implementation in C++ on CPU is significantly superior, both in terms of wall-clock time and accuracy, compared to the most optimized TensorFlow implementations of other popular negative sampling approaches on powerful NVIDIA V100 GPU.

arXiv:2012.14336

Spatiotemporal Pattern Mining for Nowcasting Extreme Earthquakes in Southern California

Geoscience and seismology have utilized the most advanced technologies and equipment to monitor seismic events globally from the past few decades. With the enormous amount of data, modern GPU-powered deep learning presents a promising approach to analyze data and discover patterns. In recent years, there are plenty of successful deep learning models for picking seismic waves. However, forecasting extreme earthquakes, which can cause disasters, is still an underdeveloped topic in history. Relevant research in spatiotemporal dynamics mining and forecasting has revealed some successful predictions, a crucial topic in many scientific research fields. Most studies of them have many successful applications of using deep neural networks. In Geology and Earth science studies, earthquake prediction is one of the world's most challenging problems, about which cutting-edge deep learning technologies may help discover some valuable patterns. In this project, we propose a deep learning modeling approach, namely \tseqpre, to mine spatiotemporal patterns from data to nowcast extreme earthquakes by discovering visual dynamics in regional coarse-grained spatial grids over time. In this modeling approach, we use synthetic deep learning neural networks with domain knowledge in geoscience and seismology to exploit earthquake patterns for prediction using convolutional long short-term memory neural networks. Our experiments show a strong correlation between location prediction and magnitude prediction for earthquakes in Southern California. Ablation studies and visualization validate the effectiveness of the proposed modeling method.

arXiv:2012.14332

Hardware-accelerated Simulation-based Inference of Stochastic Epidemiology Models for COVID-19

Epidemiology models are central in understanding and controlling large scale pandemics. Several epidemiology models require simulation-based inference such as Approximate Bayesian Computation (ABC) to fit their parameters to observations. ABC inference is highly amenable to efficient hardware acceleration. In this work, we develop parallel ABC inference of a stochastic epidemiology model for COVID-19. The statistical inference framework is implemented and compared on Intel Xeon CPU, NVIDIA Tesla V100 GPU and the Graphcore Mk1 IPU, and the results are discussed in the context of their computational architectures. Results show that GPUs are 4x and IPUs are 30x faster than Xeon CPUs. Extensive performance analysis indicates that the difference between IPU and GPU can be attributed to higher communication bandwidth, closeness of memory to compute, and higher compute power in the IPU. The proposed framework scales across 16 IPUs, with scaling overhead not exceeding 8% for the experiments performed. We present an example of our framework in practice, performing inference on the epidemiology model across three countries, and giving a brief overview of the results.

arXiv:2012.13846

SparsePipe: Parallel Deep Learning for 3D Point Clouds

We propose SparsePipe, an efficient and asynchronous parallelism approach for handling 3D point clouds with multi-GPU training. SparsePipe is built to support 3D sparse data such as point clouds. It achieves this by adopting generalized convolutions with sparse tensor representation to build expressive high-dimensional convolutional neural networks. Compared to dense solutions, the new models can efficiently process irregular point clouds without densely sliding over the entire space, significantly reducing the memory requirements and allowing higher resolutions of the underlying 3D volumes for better performance. SparsePipe exploits intra-batch parallelism that partitions input data into multiple processors and further improves the training throughput with inter-batch pipelining to overlap communication and computing. Besides, it suitably partitions the model when the GPUs are heterogeneous such that the computing is load-balanced with reduced communication overhead. Using experimental results on an eight-GPU platform, we show that SparsePipe can parallelize effectively and obtain better performance on current point cloud benchmarks for both training and inference, compared to its dense solutions.

arXiv:2012.09132

Ensemble-CVDNet: A Deep Learning based End-to-End Classification Framework for COVID-19 Detection using Ensembles of Networks

The new type of coronavirus disease (COVID-19), which started in Wuhan, China in December 2019, continues to spread rapidly affecting the whole world. It is essential to have a highly sensitive diagnostic screening tool to detect the disease as early as possible. Currently, chest CT imaging is preferred as the primary screening tool for evaluating the COVID-19 pneumonia by radiological imaging. However, CT imaging requires larger radiation doses, longer exposure time, higher cost, and may suffer from patient movements. X-Ray imaging is a fast, cheap, more patient-friendly and available in almost every healthcare facility. Therefore, we have focused on X-Ray images and developed an end-to-end deep learning model, i.e. Ensemble-CVDNet, to distinguish COVID-19 pneumonia from non-COVID pneumonia and healthy cases in this work. The proposed model is based on a combination of three lightweight pre-trained models SqueezeNet, ShuffleNet, and EfficientNet-B0 at different depths, and combines feature maps in different abstraction levels. In the proposed end to-end model, networks are used as feature extractors in parallel after fine-tuning, and some additional layers are used at the top of them. The proposed model is evaluated in the COVID-19 Radiography Database, a public data set consisting of 219 COVID-19, 1341 Healthy, and 1345 Viral Pneumonia chest X-Ray images. Experimental results show that our lightweight Ensemble-CVDNet model provides 98.30% accuracy, 97.78% sensitivity, and 97.61% F1 score using only 5.62M parameters. Moreover, it takes about 10ms to process and predict an X-Ray image using the proposed method using a mid level GPU. We believe that the method proposed in this study can be a helpful diagnostic screening tool for radiologists in the early diagnosis of the disease.

arXiv:2012.13212공식 코드 · Papers with Code

Memory-Efficient Hierarchical Neural Architecture Search for Image Restoration

Recently, much attention has been spent on neural architecture search (NAS), aiming to outperform those manually-designed neural architectures on high-level vision recognition tasks. Inspired by the success, here we attempt to leverage NAS techniques to automatically design efficient network architectures for low-level image restoration tasks. In particular, we propose a memory-efficient hierarchical NAS (termed HiNAS) and apply it to two such tasks: image denoising and image super-resolution. HiNAS adopts gradient based search strategies and builds a flexible hierarchical search space, including the inner search space and outer search space. They are in charge of designing cell architectures and deciding cell widths, respectively. For the inner search space, we propose a layer-wise architecture sharing strategy (LWAS), resulting in more flexible architectures and better performance. For the outer search space, we design a cell-sharing strategy to save memory, and considerably accelerate the search speed. The proposed HiNAS method is both memory and computation efficient. With a single GTX1080Ti GPU, it takes only about 1 hour for searching for denoising network on the BSD-500 dataset and 3.5 hours for searching for the super-resolution structure on the DIV2K dataset. Experiments show that the architectures found by HiNAS have fewer parameters and enjoy a faster inference speed, while achieving highly competitive performance compared with state-of-the-art methods. Code is available at: https://github.com/hkzhang91/HiNAS

arXiv:2012.12843

EQ-Net: A Unified Deep Learning Framework for Log-Likelihood Ratio Estimation and Quantization

In this work, we introduce EQ-Net: the first holistic framework that solves both the tasks of log-likelihood ratio (LLR) estimation and quantization using a data-driven method. We motivate our approach with theoretical insights on two practical estimation algorithms at the ends of the complexity spectrum and reveal a connection between the complexity of an algorithm and the information bottleneck method: simpler algorithms admit smaller bottlenecks when representing their solution. This motivates us to propose a two-stage algorithm that uses LLR compression as a pretext task for estimation and is focused on low-latency, high-performance implementations via deep neural networks. We carry out extensive experimental evaluation and demonstrate that our single architecture achieves state-of-the-art results on both tasks when compared to previous methods, with gains in quantization efficiency as high as $20\%$ and reduced estimation latency by up to $60\%$ when measured on general purpose and graphical processing units (GPU). In particular, our approach reduces the GPU inference latency by more than two times in several multiple-input multiple-output (MIMO) configurations. Finally, we demonstrate that our scheme is robust to distributional shifts and retains a significant part of its performance when evaluated on 5G channel models, as well as channel estimation errors.

arXiv:2012.12544

BaPipe: Exploration of Balanced Pipeline Parallelism for DNN Training

The size of deep neural networks (DNNs) grows rapidly as the complexity of the machine learning algorithm increases. To satisfy the requirement of computation and memory of DNN training, distributed deep learning based on model parallelism has been widely recognized. We propose a new pipeline parallelism training framework, BaPipe, which can automatically explore pipeline parallelism training methods and balanced partition strategies for DNN distributed training. In BaPipe, each accelerator calculates the forward propagation and backward propagation of different parts of networks to implement the intra-batch pipeline parallelism strategy. BaPipe uses a new load balancing automatic exploration strategy that considers the parameters of DNN models and the computation, memory, and communication resources of accelerator clusters. We have trained different DNNs such as VGG-16, ResNet-50, and GNMT on GPU clusters and simulated the performance of different FPGA clusters. Compared with state-of-the-art data parallelism and pipeline parallelism frameworks, BaPipe provides up to 3.2x speedup and 4x memory reduction in various platforms.

arXiv:2012.12540공식 코드 · Papers with Code

Evolving Neural Architecture Using One Shot Model

Neural Architecture Search (NAS) is emerging as a new research direction which has the potential to replace the hand-crafted neural architectures designed for specific tasks. Previous evolution based architecture search requires high computational resources resulting in high search time. In this work, we propose a novel way of applying a simple genetic algorithm to the NAS problem called EvNAS (Evolving Neural Architecture using One Shot Model) which reduces the search time significantly while still achieving better result than previous evolution based methods. The architectures are represented by using the architecture parameter of the one shot model which results in the weight sharing among the architectures for a given population of architectures and also weight inheritance from one generation to the next generation of architectures. We propose a decoding technique for the architecture parameter which is used to divert majority of the gradient information towards the given architecture and is also used for improving the performance prediction of the given architecture from the one shot model during the search process. Furthermore, we use the accuracy of the partially trained architecture on the validation data as a prediction of its fitness in order to reduce the search time. EvNAS searches for the architecture on the proxy dataset i.e. CIFAR-10 for 4.4 GPU day on a single GPU and achieves top-1 test error of 2.47% with 3.63M parameters which is then transferred to CIFAR-100 and ImageNet achieving top-1 error of 16.37% and top-5 error of 7.4% respectively. All of these results show the potential of evolutionary methods in solving the architecture search problem.

arXiv:2012.11225

Searching for Controllable Image Restoration Networks

Diverse user preferences over images have recently led to a great amount of interest in controlling the imagery effects for image restoration tasks. However, existing methods require separate inference through the entire network per each output, which hinders users from readily comparing multiple imagery effects due to long latency. To this end, we propose a novel framework based on a neural architecture search technique that enables efficient generation of multiple imagery effects via two stages of pruning: task-agnostic and task-specific pruning. Specifically, task-specific pruning learns to adaptively remove the irrelevant network parameters for each task, while task-agnostic pruning learns to find an efficient architecture by sharing the early layers of the network across different tasks. Since the shared layers allow for feature reuse, only a single inference of the task-agnostic layers is needed to generate multiple imagery effects from the input image. Using the proposed task-agnostic and task-specific pruning schemes together significantly reduces the FLOPs and the actual latency of inference compared to the baseline. We reduce 95.7% of the FLOPs when generating 27 imagery effects, and make the GPU latency 73.0% faster on 4K-resolution images.

arXiv:2012.10802

Rethinking Road Surface 3D Reconstruction and Pothole Detection: From Perspective Transformation to Disparity Map Segmentation

Potholes are one of the most common forms of road damage, which can severely affect driving comfort, road safety and vehicle condition. Pothole detection is typically performed by either structural engineers or certified inspectors. This task is, however, not only hazardous for the personnel but also extremely time-consuming. This paper presents an efficient pothole detection algorithm based on road disparity map estimation and segmentation. We first generalize the perspective transformation by incorporating the stereo rig roll angle. The road disparities are then estimated using semi-global matching. A disparity map transformation algorithm is then performed to better distinguish the damaged road areas. Finally, we utilize simple linear iterative clustering to group the transformed disparities into a collection of superpixels. The potholes are then detected by finding the superpixels, whose values are lower than an adaptively determined threshold. The proposed algorithm is implemented on an NVIDIA RTX 2080 Ti GPU in CUDA. The experiments demonstrate the accuracy and efficiency of our proposed road pothole detection algorithm, where an accuracy of 99.6% and an F-score of 89.4% are achieved.