國立虎尾科技大學 |

Caffeinated FPGAs : = FPGA Framework for Training and Inference of Convolutional Neural Networks With Reduced Precision Floating-Point Arithmetic.

紀錄類型:	書目-語言資料,手稿 : Monograph/item
正題名/作者:	Caffeinated FPGAs :/
其他題名:	FPGA Framework for Training and Inference of Convolutional Neural Networks With Reduced Precision Floating-Point Arithmetic.
作者:	DiCecco, Roberto.
面頁冊數:	1 online resource (87 pages)
附註:	Source: Masters Abstracts International, Volume: 58-01.
Contained By:	Masters Abstracts International58-01(E).
標題:	Computer engineering. -
電子資源:	click for full text (PQDT)
ISBN:	9780438185005

Caffeinated FPGAs : = FPGA Framework for Training and Inference of Convolutional Neural Networks With Reduced Precision Floating-Point Arithmetic.
DiCecco, Roberto.

Caffeinated FPGAs :FPGA Framework for Training and Inference of Convolutional Neural Networks With Reduced Precision Floating-Point Arithmetic. - 1 online resource (87 pages)

Source: Masters Abstracts International, Volume: 58-01.

Thesis (M.A.S.)--University of Toronto (Canada), 2018.

Includes bibliographical references

This thesis presents a framework for performing training and inference of Convolutional Neural Networks (CNNs) with reduced precision floating-point arithmetic. This work aims to provide a means for FPGA and machine learning researchers to use the customizability of FPGAs to explore the precision requirements of training CNNs with an open-source framework. This is accomplished through the creation of a High-Level Synthesis library with a Custom Precision Floating-Point data type that is configurable in both exponent and mantissa widths, with several standard operators and rounding modes supported. With this library a FPGA CNN Training Engine (FCTE) has been created along with a FPGA CNN framework FPGA Caffe, which is built on Caffe. FCTE has a peak performance of approximately 350 GFLOPs, and has been used to show that a mantissa width of 5 and exponent width of 6 is sufficient for training several models targeting the MNIST and CIFAR-10 datasets.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2018

Mode of access: World Wide Web

ISBN: 9780438185005Subjects--Topical Terms:

569006
Computer engineering.
Index Terms--Genre/Form:

554714
Electronic books.

Caffeinated FPGAs : = FPGA Framework for Training and Inference of Convolutional Neural Networks With Reduced Precision Floating-Point Arithmetic.
LDR:02237ntm a2200337Ki 4500 001 916894
005 20180928111502.5
006 m o u
007 cr mn||||a|a||
008 190606s2018 xx obm 000 0 eng d
020 $a 9780438185005
035 $a (MiAaPQ)AAI10791098
035 $a (MiAaPQ)toronto:17464
035 $a AAI10791098
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a DiCecco, Roberto. $3 1190758
245 1 0 $a Caffeinated FPGAs : $b FPGA Framework for Training and Inference of Convolutional Neural Networks With Reduced Precision Floating-Point Arithmetic.
264 0 $c 2018
300 $a 1 online resource (87 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: Masters Abstracts International, Volume: 58-01.
500 $a Adviser: Paul Chow.
502 $a Thesis (M.A.S.)--University of Toronto (Canada), 2018.
504 $a Includes bibliographical references
520 $a This thesis presents a framework for performing training and inference of Convolutional Neural Networks (CNNs) with reduced precision floating-point arithmetic. This work aims to provide a means for FPGA and machine learning researchers to use the customizability of FPGAs to explore the precision requirements of training CNNs with an open-source framework. This is accomplished through the creation of a High-Level Synthesis library with a Custom Precision Floating-Point data type that is configurable in both exponent and mantissa widths, with several standard operators and rounding modes supported. With this library a FPGA CNN Training Engine (FCTE) has been created along with a FPGA CNN framework FPGA Caffe, which is built on Caffe. FCTE has a peak performance of approximately 350 GFLOPs, and has been used to show that a mantissa width of 5 and exponent width of 6 is sufficient for training several models targeting the MNIST and CIFAR-10 datasets.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2018
538 $a Mode of access: World Wide Web
650 4 $a Computer engineering. $3 569006
650 4 $a Artificial intelligence. $3 559380
655 7 $a Electronic books. $2 local $3 554714
690 $a 0464
690 $a 0800
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a University of Toronto (Canada). $b Electrical and Computer Engineering. $3 1148628
773 0 $t Masters Abstracts International $g 58-01(E).
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10791098 $z click for full text (PQDT)