Graph convolutional networks (GCNs) have been introduced to effectively process non-euclidean graph data. However, GCNs incur large amounts of irregularity in computation and memory access, which prevents efficient use of traditional neural network accelerators. Moreover, existing dedicated GCN accelerators demand high memory volumes and are difficult to implement onto resource limited edge devices. In this work, we propose LW-GCN, a lightweight FPGA-based accelerator with a software-hardware co-designed process to tackle irregularity in computation and memory access in GCN inference. LW-GCN decomposes the main GCN operations into sparse-dense matrix multiplication (SDMM) and dense matrix multiplication (DMM). We propose a novel compression format to balance workload across PEs and prevent data hazards. Moreover, we apply data quantization and workload tiling, and map both SDMM and DMM of GCN inference onto a uniform architecture on resource limited hardware. Evaluation on GCN and GraphSAGE are performed on Xilinx Kintex-7 FPGA with three popular datasets. Compared to existing CPU, GPU, and state-of-the-art FPGA-based accelerator, LW-GCN reduces latency by up to 60x, 12x and 1.7x and increases power efficiency by up to 912x., 511x and 3.87x, respectively. Furthermore, compared with NVIDIA's latest edge GPU Jetson Xavier NX, LW-GCN achieves speedup and energy savings of 32x and 84x, respectively.
Tea chrysanthemum detection at its flowering stage is one of the key components for selective chrysanthemum harvesting robot development. However, it is a challenge to detect flowering chrysanthemums under unstructured field environments given the variations on illumination, occlusion and object scale. In this context, we propose a highly fused and lightweight deep learning architecture based on YOLO for tea chrysanthemum detection (TC-YOLO). First, in the backbone component and neck component, the method uses the Cross-Stage Partially Dense Network (CSPDenseNet) as the main network, and embeds custom feature fusion modules to guide the gradient flow. In the final head component, the method combines the recursive feature pyramid (RFP) multiscale fusion reflow structure and the Atrous Spatial Pyramid Pool (ASPP) module with cavity convolution to achieve the detection task. The resulting model was tested on 300 field images, showing that under the NVIDIA Tesla P100 GPU environment, if the inference speed is 47.23 FPS for each image (416 * 416), TC-YOLO can achieve the average precision (AP) of 92.49% on our own tea chrysanthemum dataset. In addition, this method (13.6M) can be deployed on a single mobile GPU, and it could be further developed as a perception system for a selective chrysanthemum harvesting robot in the future.
Whole abdominal organ segmentation is important in diagnosing abdomen lesions, radiotherapy, and follow-up. However, oncologists' delineating all abdominal organs from 3D volumes is time-consuming and very expensive. Deep learning-based medical image segmentation has shown the potential to reduce manual delineation efforts, but it still requires a large-scale fine annotated dataset for training, and there is a lack of large-scale datasets covering the whole abdomen region with accurate and detailed annotations for the whole abdominal organ segmentation. In this work, we establish a new large-scale \textit{W}hole abdominal \textit{OR}gan \textit{D}ataset (\textit{WORD}) for algorithm research and clinical application development. This dataset contains 150 abdominal CT volumes (30495 slices). Each volume has 16 organs with fine pixel-level annotations and scribble-based sparse annotations, which may be the largest dataset with whole abdominal organ annotation. Several state-of-the-art segmentation methods are evaluated on this dataset. And we also invited three experienced oncologists to revise the model predictions to measure the gap between the deep learning method and oncologists. Afterwards, we investigate the inference-efficient learning on the WORD, as the high-resolution image requires large GPU memory and a long inference time in the test stage. We further evaluate the scribble-based annotation-efficient learning on this dataset, as the pixel-wise manual annotation is time-consuming and expensive. The work provided a new benchmark for the abdominal multi-organ segmentation task, and these experiments can serve as the baseline for future research and clinical application development.
We study how to support elasticity, that is, the ability to dynamically adjust the parallelism (i.e., the number of GPUs), for deep neural network (DNN) training in a GPU cluster. Elasticity can benefit multi-tenant GPU cluster management in many ways, for example, achieving various scheduling objectives (e.g., job throughput, job completion time, GPU efficiency) according to cluster load variations, utilizing transient idle resources, and supporting performance profiling, job migration, and straggler mitigation. We propose EDL, which enables elastic deep learning with a simple API and can be easily integrated with existing deep learning frameworks such as TensorFlow and PyTorch. EDL also incorporates techniques that are necessary to reduce the overhead of parallelism adjustments, such as stop-free scaling and dynamic data pipeline. We demonstrate with experiments that EDL can indeed bring significant benefits to the above-listed applications in GPU cluster management.
Optimizing compilers exploit the memory hierarchy using loop tiling and fusion, but these two transformations usually interfere with each other due to the oversight of transformations on data in memories. We present a novel composition of loop tiling and fusion in this paper. Unlike existing tiling-after-fusion algorithms that only transform computation spaces, our approach first applies rectangular/parallelogram tiling to live-out computation spaces for fitting the memory hierarchy, followed by the computation of the memory footprints required by each tile. The upwards exposed data extracted from the memory footprints are used to determine the tile shapes of intermediate computation spaces, allowing the construction of arbitrary tile shapes. Finally, our technique implements a post-tiling fusion strategy for maximizing data locality without losing tilability or parallelism of live-out computation spaces, thereby enabling storage reduction and reuse, and optimizing the memory hierarchy. We demonstrate that our approach can achieve superior performance on both CPU and GPU architectures over the state of the art by experimenting on 11 benchmarks extracted from numerous domains including neural networks, image processing, sparse matrix computation and linear algebra. Also, the results of the ResNet-50 model on an AI accelerator show that our approach can obtain 16% performance improvement.
Emerging graph neural networks (GNNs) have extended the successes of deep learning techniques against datasets like images and texts to more complex graph-structured data. By leveraging GPU accelerators, existing frameworks combine both mini-batch and sampling for effective and efficient model training on large graphs. However, this setup faces a scalability issue since loading rich vertices features from CPU to GPU through a limited bandwidth link usually dominates the training cycle. In this paper, we propose PaGraph, a system that supports general and efficient sampling-based GNN training on single-server with multi-GPU. PaGraph significantly reduces the data loading time by exploiting available GPU resources to keep frequently accessed graph data with a cache. It also embodies a lightweight yet effective caching policy that takes into account graph structural information and data access patterns of sampling-based GNN training simultaneously. Furthermore, to scale out on multiple GPUs, PaGraph develops a fast GNN-computation-aware partition algorithm to avoid cross-partition access during data parallel training and achieves better cache efficiency. Evaluations on two representative GNN models, GCN and GraphSAGE, show that PaGraph achieves up to 96.8% data loading time reductions and up to 4.8X performance speedup over the state-of-the-art baselines. Together with preprocessing optimization, PaGraph further delivers up to 16.0X end-to-end speedup.
Graph neural networks (GNNs) have achieved breakthrough performance in graph analytics such as node classification, link prediction and graph clustering. Many GNN training frameworks have been developed, but they are usually designed as a set of manually written, GNN-specific operators plugged into existing deep learning systems, which incurs high memory consumption, poor data locality, and large semantic gap between algorithm design and implementation. This paper proposes the Seastar system, which presents a vertex-centric programming model for GNN training on GPU and provides idiomatic python constructs to enable easy development of novel homogeneous and heterogeneous GNN models. We also propose novel optimizations to produce highly efficient fused GPU kernels for forward and backward passes in GNN training. Compared with the state-of-the art GNN systems, DGL and PyG, Seastar achieves better usability, up to 2 and 8 times less memory consumption, and 14 and 3 times faster execution, respectively.
Existing tensor compilers have proven their effectiveness in deploying deep neural networks on general-purpose hardware like CPU and GPU, but optimizing for neural processing units (NPUs) is still challenging due to the heterogeneous compute units and complicated memory hierarchy. In this paper, we present AKG, a tensor compiler for NPUs. AKG first lowers the tensor expression language to a polyhedral representation, which is used to automate the memory management of NPUs. Unlike existing approaches that resort to manually written schedules, AKG leverages polyhedral schedulers to perform a much wider class of transformations, and extends the semantics of the polyhedral representation to combine complex tiling techniques and hierarchical fusion strategies. We also implement the domain-specific optimization of convolution in AKG. Moreover, to achieve the optimal performance, we introduce complementary optimizations in code generation, which is followed by an auto-tuner. We conduct extensive experiments on benchmarks ranging from single operators to end-to-end networks. The experimental results show that AKG can obtain superior performance to both manual scheduling approaches and vendor provided libraries. We believe AKG will cast a light on the follow-up compiler works on NPUs.
We propose an accurate and efficient scene text detection framework, termed FAST (i.e., faster arbitrarily-shaped text detector). Different from recent advanced text detectors that used complicated post-processing and hand-crafted network architectures, resulting in low inference speed, FAST has two new designs. (1) We design a minimalist kernel representation (only has 1-channel output) to model text with arbitrary shape, as well as a GPU-parallel post-processing to efficiently assemble text lines with a negligible time overhead. (2) We search the network architecture tailored for text detection, leading to more powerful features than most networks that are searched for image classification. Benefiting from these two designs, FAST achieves an excellent trade-off between accuracy and efficiency on several challenging datasets, including Total Text, CTW1500, ICDAR 2015, and MSRA-TD500. For example, FAST-T yields 81.6% F-measure at 152 FPS on Total-Text, outperforming the previous fastest method by 1.7 points and 70 FPS in terms of accuracy and speed. With TensorRT optimization, the inference speed can be further accelerated to over 600 FPS. Code and models will be released at https://github.com/czczup/FAST.
This paper presents a comparison of various deep learning models such as Exception, Resnet-50, and Inception V3 for the classification and prediction of hand hygiene gestures, which were recorded in accordance with the World Health Organization (WHO) guidelines. The dataset consists of six hand hygiene movements in a video format, gathered for 30 participants. The network consists of pre-trained models with image net weights and a modified head of the model. An accuracy of 37% (Xception model), 33% (Inception V3), and 72% (ResNet-50) is achieved in the classification report after the training of the models for 25 epochs. ResNet-50 model clearly outperforms with correct class predictions. The major speed limitation can be overcome with the use of fast processing GPU for future work. A complete hand hygiene dataset along with other generic gestures such as one-hand movements (linear hand motion; circular hand rotation) will be tested with ResNet-50 architecture and the variants for health care workers.
Deep Q-Network (DQN) marked a major milestone for reinforcement learning, demonstrating for the first time that human-level control policies could be learned directly from raw visual inputs via reward maximization. Even years after its introduction, DQN remains highly relevant to the research community since many of its innovations have been adopted by successor methods. Nevertheless, despite significant hardware advances in the interim, DQN's original Atari 2600 experiments remain costly to replicate in full. This poses an immense barrier to researchers who cannot afford state-of-the-art hardware or lack access to large-scale cloud computing resources. To facilitate improved access to deep reinforcement learning research, we introduce a DQN implementation that leverages a novel concurrent and synchronized execution framework designed to maximally utilize a heterogeneous CPU-GPU desktop system. With just one NVIDIA GeForce GTX 1080 GPU, our implementation reduces the training time of a 200-million-frame Atari experiment from 25 hours to just 9 hours. The ideas introduced in our paper should be generalizable to a large number of off-policy deep reinforcement learning methods.
Recently, Graph Neural Networks (GNNs) have become state-of-the-art algorithms for analyzing non-euclidean graph data. However, to realize efficient GNN training is challenging, especially on large graphs. The reasons are many-folded: 1) GNN training incurs a substantial memory footprint. Full-batch training on large graphs even requires hundreds to thousands of gigabytes of memory. 2) GNN training involves both memory-intensive and computation-intensive operations, challenging current CPU/GPU platforms. 3) The irregularity of graphs can result in severe resource under-utilization and load-imbalance problems. This paper presents a GNNear accelerator to tackle these challenges. GNNear adopts a DIMM-based memory system to provide sufficient memory capacity. To match the heterogeneous nature of GNN training, we offload the memory-intensive Reduce operations to in-DIMM Near-Memory-Engines (NMEs), making full use of the high aggregated local bandwidth. We adopt a Centralized-Acceleration-Engine (CAE) to process the computation-intensive Update operations. We further propose several optimization strategies to deal with the irregularity of input graphs and improve GNNear's performance. Comprehensive evaluations on 16 GNN training tasks demonstrate that GNNear achieves 30.8$\times$/2.5$\times$ geomean speedup and 79.6$\times$/7.3$\times$(geomean) higher energy efficiency compared to Xeon E5-2698-v4 CPU and NVIDIA V100 GPU.
The strong demand for efficient and performant deployment of Deep Learning (DL) applications prompts the rapid development of a rich DL ecosystem. To keep up with this fast advancement, it is crucial for modern DL frameworks to efficiently integrate a variety of optimized tensor algebra libraries and runtimes as their backends and generate the fastest possible executable using these backends. However, current DL frameworks require significant manual effort and expertise to integrate every new backend while failing to unleash its full potential. Given the fast-evolving nature of the DL ecosystem, this manual approach often slows down continuous innovations across different layers; it prevents hardware vendors from the fast deployment of their cutting-edge libraries, DL framework developers must repeatedly adjust their hand-coded rules to accommodate new versions of libraries, and machine learning practitioners need to wait for the integration of new technologies and often encounter unsatisfactory performance. In this paper, we propose Collage, a DL framework that offers seamless integration of DL backends. Collage provides an expressive backend registration interface that allows users to precisely specify the capability of various backends. By leveraging the specifications of available backends, Collage automatically searches for an optimized backend placement strategy for a given workload and execution environment. Our evaluation shows that Collage outperforms the best existing framework for each hardware by $1.26\times$, $1.43\times$, $1.40\times$ on average on NVIDIA's RTX 2070 GPU, V100 GPU, and Intel's Xeon 8259CL CPU, respectively. Collage has been open-sourced and deployed in Apache TVM.
For its advantage in GPU acceleration and less dependency on human experts, machine learning has been an emerging tool for solving the placement and routing problems, as two critical steps in modern chip design flow. Being still in its early stage, there are fundamental issues: scalability, reward design, and end-to-end learning paradigm etc. To achieve end-to-end placement learning, we first propose a joint learning method termed by DeepPlace for the placement of macros and standard cells, by the integration of reinforcement learning with a gradient based optimization scheme. To further bridge the placement with the subsequent routing task, we also develop a joint learning approach via reinforcement learning to fulfill both macro placement and routing, which is called DeepPR. One key design in our (reinforcement) learning paradigm involves a multi-view embedding model to encode both global graph level and local node level information of the input macros. Moreover, the random network distillation is devised to encourage exploration. Experiments on public chip design benchmarks show that our method can effectively learn from experience and also provides intermediate placement for the post standard cell placement, within few hours for training.
Most research on novel techniques for 3D Medical Image Segmentation (MIS) is currently done using Deep Learning with GPU accelerators. The principal challenge of such technique is that a single input can easily cope computing resources, and require prohibitive amounts of time to be processed. Distribution of deep learning and scalability over computing devices is an actual need for progressing on such research field. Conventional distribution of neural networks consist in data parallelism, where data is scattered over resources (e.g., GPUs) to parallelize the training of the model. However, experiment parallelism is also an option, where different training processes are parallelized across resources. While the first option is much more common on 3D image segmentation, the second provides a pipeline design with less dependence among parallelized processes, allowing overhead reduction and more potential scalability. In this work we present a design for distributed deep learning training pipelines, focusing on multi-node and multi-GPU environments, where the two different distribution approaches are deployed and benchmarked. We take as proof of concept the 3D U-Net architecture, using the MSD Brain Tumor Segmentation dataset, a state-of-art problem in medical image segmentation with high computing and space requirements. Using the BSC MareNostrum supercomputer as benchmarking environment, we use TensorFlow and Ray as neural network training and experiment distribution platforms. We evaluate the experiment speed-up, showing the potential for scaling out on GPUs and nodes. Also comparing the different parallelism techniques, showing how experiment distribution leverages better such resources through scaling. Finally, we provide the implementation of the design open to the community, and the non-trivial steps and methodology for adapting and deploying a MIS case as the here presented.
This document describes version 0.10 of TorchAudio: building blocks for machine learning applications in the audio and speech processing domain. The objective of TorchAudio is to accelerate the development and deployment of machine learning applications for researchers and engineers by providing off-the-shelf building blocks. The building blocks are designed to be GPU-compatible, automatically differentiable, and production-ready. TorchAudio can be easily installed from Python Package Index repository and the source code is publicly available under a BSD-2-Clause License (as of September 2021) at https://github.com/pytorch/audio. In this document, we provide an overview of the design principles, functionalities, and benchmarks of TorchAudio. We also benchmark our implementation of several audio and speech operations and models. We verify through the benchmarks that our implementations of various operations and models are valid and perform similarly to other publicly available implementations.
This paper presents a novel approach for background/foreground segmentation of RGBD data with the Gaussian Mixture Models (GMM). We first start by the background subtraction from the colour and depth images separately. The foregrounds resulting from both streams are then fused for a more accurate detection. Our segmentation solution is implemented on the GPU. Thus, it works at the full frame rate of the sensor (30fps). Test results show its robustness against illumination change, shadows and reflections.
Accurate segmentation is a crucial step in medical image analysis and applying supervised machine learning to segment the organs or lesions has been substantiated effective. However, it is costly to perform data annotation that provides ground truth labels for training the supervised algorithms, and the high variance of data that comes from different domains tends to severely degrade system performance over cross-site or cross-modality datasets. To mitigate this problem, a novel unsupervised domain adaptation (UDA) method named dispensed Transformer network (DTNet) is introduced in this paper. Our novel DTNet contains three modules. First, a dispensed residual transformer block is designed, which realizes global attention by dispensed interleaving operation and deals with the excessive computational cost and GPU memory usage of the Transformer. Second, a multi-scale consistency regularization is proposed to alleviate the loss of details in the low-resolution output for better feature alignment. Finally, a feature ranking discriminator is introduced to automatically assign different weights to domain-gap features to lessen the feature distribution distance, reducing the performance shift of two domains. The proposed method is evaluated on large fluorescein angiography (FA) retinal nonperfusion (RNP) cross-site dataset with 676 images and a wide used cross-modality dataset from the MM-WHS challenge. Extensive results demonstrate that our proposed network achieves the best performance in comparison with several state-of-the-art techniques.
Knowledge graphs (KGs) capture knowledge in the form of head--relation--tail triples and are a crucial component in many AI systems. There are two important reasoning tasks on KGs: (1) single-hop knowledge graph completion, which involves predicting individual links in the KG; and (2), multi-hop reasoning, where the goal is to predict which KG entities satisfy a given logical query. Embedding-based methods solve both tasks by first computing an embedding for each entity and relation, then using them to form predictions. However, existing scalable KG embedding frameworks only support single-hop knowledge graph completion and cannot be applied to the more challenging multi-hop reasoning task. Here we present Scalable Multi-hOp REasoning (SMORE), the first general framework for both single-hop and multi-hop reasoning in KGs. Using a single machine SMORE can perform multi-hop reasoning in Freebase KG (86M entities, 338M edges), which is 1,500x larger than previously considered KGs. The key to SMORE's runtime performance is a novel bidirectional rejection sampling that achieves a square root reduction of the complexity of online training data generation. Furthermore, SMORE exploits asynchronous scheduling, overlapping CPU-based data sampling, GPU-based embedding computation, and frequent CPU--GPU IO. SMORE increases throughput (i.e., training speed) over prior multi-hop KG frameworks by 2.2x with minimal GPU memory requirements (2GB for training 400-dim embeddings on 86M-node Freebase) and achieves near linear speed-up with the number of GPUs. Moreover, on the simpler single-hop knowledge graph completion task SMORE achieves comparable or even better runtime performance to state-of-the-art frameworks on both single GPU and multi-GPU settings.
The success of Transformer models has pushed the deep learning model scale to billions of parameters. Due to the limited memory resource of a single GPU, However, the best practice for choosing the optimal parallel strategy is still lacking, since it requires domain expertise in both deep learning and parallel computing. The Colossal-AI system addressed the above challenge by introducing a unified interface to scale your sequential code of model training to distributed environments. It supports parallel training methods such as data, pipeline, tensor, and sequence parallelism, as well as heterogeneous training methods integrated with zero redundancy optimizer. Compared to the baseline system, Colossal-AI can achieve up to 2.76 times training speedup on large-scale models.