國立虎尾科技大學 |

End-to-End Inference Optimization for Deep Learning-Based Image Upsampling Networks.

紀錄類型:	書目-語言資料,手稿 : Monograph/item
正題名/作者:	End-to-End Inference Optimization for Deep Learning-Based Image Upsampling Networks./
作者:	Colbert, Ian.
面頁冊數:	1 online resource (175 pages)
附註:	Source: Dissertations Abstracts International, Volume: 85-01, Section: B.
Contained By:	Dissertations Abstracts International85-01B.
標題:	Electrical engineering. -
電子資源:	click for full text (PQDT)
ISBN:	9798379796150

End-to-End Inference Optimization for Deep Learning-Based Image Upsampling Networks.
Colbert, Ian.

End-to-End Inference Optimization for Deep Learning-Based Image Upsampling Networks. - 1 online resource (175 pages)

Source: Dissertations Abstracts International, Volume: 85-01, Section: B.

Thesis (Ph.D.)--University of California, San Diego, 2023.

Includes bibliographical references

Many computer vision problems require image upsampling, where the number of pixels per unit area is increased by inferring values in high-dimensional image space from lowdimensional representations. Recent research has shown that deep learning-based solutions achieve state-of-the-art performance on such tasks by training deep neural networks (DNNs) on large annotated datasets. Yet, their adoption in real-time applications is predicated on the deployment costs of the resulting models since end-user devices impose significant compute and memory constraints on inference pipelines. To address this, many researchers and practitioners have proposed methods to reduce inference costs without sacrificing model quality. However, many of these works focus on DNNs designed for image downsampling. In this thesis, we study inference optimization techniques designed for deep learning-based image upsampling networks. While some inference optimizations are applicable to both upsampling and downsampling networks, we show that specifically tailoring optimizations for image upsampling workloads can lead to more efficient and effective deployment.We maintain a holistic view of inference optimization, from training through deployment to execution, by integrating hardware-aware deep learning techniques, compute graph transformations, and computer architecture optimizations into an end-to-end pipeline. We begin by characterizing this pipeline and the different requirements for image upsampling and downsampling workloads. We then introduce novel statistical approaches to hardware-aware deep learning techniques based on quantization and pruning. Once trained, we then introduce novel compute kernels and graph transformations that reduce the compute costs of common upsampling workloads by up to a factor of 3.3. Finally, we adapt our novel inference algorithms to a specialized hardware architecture that reduces resource utilization and improves dataflow on FPGA-based accelerators.We evaluate a wide range of computer vision benchmarks covering both stochastic and deterministic models to show that our approaches improve power efficiency, throughput, and resource utilization without damaging model quality. Our research highlights the importance of end-to-end inference optimization for deep learning-based image upsampling networks and provides an effective solution for reducing the deployment costs of DNNs designed for real-time computer vision applications on resource-constrained platforms.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2024

Mode of access: World Wide Web

ISBN: 9798379796150Subjects--Topical Terms:

596380
Electrical engineering.
Subjects--Index Terms:

Deep neural networksIndex Terms--Genre/Form:

554714
Electronic books.

End-to-End Inference Optimization for Deep Learning-Based Image Upsampling Networks.
LDR:03935ntm a22004097 4500 001 1143775
005 20240517104949.5
006 m o d
007 cr mn ---uuuuu
008 250605s2023 xx obm 000 0 eng d
020 $a 9798379796150
035 $a (MiAaPQ)AAI30486690
035 $a AAI30486690
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Colbert, Ian. $3 1468566
245 1 0 $a End-to-End Inference Optimization for Deep Learning-Based Image Upsampling Networks.
264 0 $c 2023
300 $a 1 online resource (175 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 85-01, Section: B.
500 $a Advisor: Kreutz-Delgado, Kenneth.
502 $a Thesis (Ph.D.)--University of California, San Diego, 2023.
504 $a Includes bibliographical references
520 $a Many computer vision problems require image upsampling, where the number of pixels per unit area is increased by inferring values in high-dimensional image space from lowdimensional representations. Recent research has shown that deep learning-based solutions achieve state-of-the-art performance on such tasks by training deep neural networks (DNNs) on large annotated datasets. Yet, their adoption in real-time applications is predicated on the deployment costs of the resulting models since end-user devices impose significant compute and memory constraints on inference pipelines. To address this, many researchers and practitioners have proposed methods to reduce inference costs without sacrificing model quality. However, many of these works focus on DNNs designed for image downsampling. In this thesis, we study inference optimization techniques designed for deep learning-based image upsampling networks. While some inference optimizations are applicable to both upsampling and downsampling networks, we show that specifically tailoring optimizations for image upsampling workloads can lead to more efficient and effective deployment.We maintain a holistic view of inference optimization, from training through deployment to execution, by integrating hardware-aware deep learning techniques, compute graph transformations, and computer architecture optimizations into an end-to-end pipeline. We begin by characterizing this pipeline and the different requirements for image upsampling and downsampling workloads. We then introduce novel statistical approaches to hardware-aware deep learning techniques based on quantization and pruning. Once trained, we then introduce novel compute kernels and graph transformations that reduce the compute costs of common upsampling workloads by up to a factor of 3.3. Finally, we adapt our novel inference algorithms to a specialized hardware architecture that reduces resource utilization and improves dataflow on FPGA-based accelerators.We evaluate a wide range of computer vision benchmarks covering both stochastic and deterministic models to show that our approaches improve power efficiency, throughput, and resource utilization without damaging model quality. Our research highlights the importance of end-to-end inference optimization for deep learning-based image upsampling networks and provides an effective solution for reducing the deployment costs of DNNs designed for real-time computer vision applications on resource-constrained platforms.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2024
538 $a Mode of access: World Wide Web
650 4 $a Electrical engineering. $3 596380
650 4 $a Computer engineering. $3 569006
650 4 $a Computer science. $3 573171
653 $a Deep neural networks
653 $a Image upsampling
653 $a Inference optimization
653 $a Machine learning
653 $a Quantization
655 7 $a Electronic books. $2 local $3 554714
690 $a 0800
690 $a 0984
690 $a 0464
690 $a 0544
710 2 $a University of California, San Diego. $b Electrical and Computer Engineering. $3 1465765
710 2 $a ProQuest Information and Learning Co. $3 1178819
773 0 $t Dissertations Abstracts International $g 85-01B.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30486690 $z click for full text (PQDT)