Compression and Distribution of a Neural Network With IoT Applications

In order to enable deployment of large neuralnetwork models on devices with limited memory capacity, refinedmethods for compressing these are essential. This project aimsat investigating some possible solutions, namely pruning andpartitioned logit based knowledge distillation, using teacherstudentle...

Full description

Bibliographic Details
Main Authors: Backe, Hannes, Rydberg, David
Format: Bachelor Thesis
Language:English
Published: KTH, Skolan för elektroteknik och datavetenskap (EECS) 2021
Subjects:
IoT
DML
Online Access:http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-308460
id ftkthstockholm:oai:DiVA.org:kth-308460
record_format openpolar
spelling ftkthstockholm:oai:DiVA.org:kth-308460 2023-05-15T16:02:06+02:00 Compression and Distribution of a Neural Network With IoT Applications Backe, Hannes Rydberg, David 2021 application/pdf http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-308460 eng eng KTH, Skolan för elektroteknik och datavetenskap (EECS) TRITA-EECS-EX 2021:186 http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-308460 info:eu-repo/semantics/openAccess Machine Learning Neural Network IoT Compression Pruning Knowledge Distillation (KD) Distributed Machine Learning (DML) Electrical Engineering Electronic Engineering Information Engineering Elektroteknik och elektronik Student thesis info:eu-repo/semantics/bachelorThesis text 2021 ftkthstockholm 2022-08-11T12:33:58Z In order to enable deployment of large neuralnetwork models on devices with limited memory capacity, refinedmethods for compressing these are essential. This project aimsat investigating some possible solutions, namely pruning andpartitioned logit based knowledge distillation, using teacherstudentlearning methods. A cumbersome benchmark teacherneural network was developed and used as a reference. A specialcase of logit based teacher-student learning was then applied,resulting not only in a compressed model, but also in a convenientway of distributing it. The individual student models were ableto mimic the parts of the teacher model with small losses, whilethe network of student models achieved similar accuracy as theteacher model. Overall, the size of the network of student modelswas around 11% of the teacher. Another popular method ofcompressing neural networks was also tested - pruning. Pruningthe teacher network resulted in a much smaller model, around18% of the teacher model, with similar accuracy. För att möjliggöra användning av storaneurala nätverksmodeller på enheter med begränsad minneskapacitetkrävs raffinerade metoder för komprimering av dessa.Detta projekt syftar till att undersöka några möjliga lösningar,nämligen pruning och partitionerad logit-baserad knowledgedistillation, med hjälp av teacher-student-träning. Ett stortriktmärkesnätverk utvecklades och användes som referens. Enspeciell typ av logit-baserad teacher-student-träning tillämpadessedan, vilket inte bara resulterade i en komprimerad modellutan också i ett smidigt sätt att distribuera den på. De enskildastudent-modellerna kunde efterlikna delar av teachermodellenmed små förluster, medan nätverket av studentmodelleruppnådde ungefär samma noggrannhet som teachermodellen.Sammantaget uppmättes storleken av nätverket avstudent-modeller till cirka 11 % av teacher-modellen. En annanpopulär metod för komprimering av neurala nätverk testadesockså pruning. Pruning av teacher-modellen resulterade i enmycket mindre modell, cirka 18 % av ... Bachelor Thesis DML Royal Institute of Technology, Stockholm: KTHs Publication Database DiVA
institution Open Polar
collection Royal Institute of Technology, Stockholm: KTHs Publication Database DiVA
op_collection_id ftkthstockholm
language English
topic Machine Learning
Neural Network
IoT
Compression
Pruning
Knowledge Distillation (KD)
Distributed Machine Learning (DML)
Electrical Engineering
Electronic Engineering
Information Engineering
Elektroteknik och elektronik
spellingShingle Machine Learning
Neural Network
IoT
Compression
Pruning
Knowledge Distillation (KD)
Distributed Machine Learning (DML)
Electrical Engineering
Electronic Engineering
Information Engineering
Elektroteknik och elektronik
Backe, Hannes
Rydberg, David
Compression and Distribution of a Neural Network With IoT Applications
topic_facet Machine Learning
Neural Network
IoT
Compression
Pruning
Knowledge Distillation (KD)
Distributed Machine Learning (DML)
Electrical Engineering
Electronic Engineering
Information Engineering
Elektroteknik och elektronik
description In order to enable deployment of large neuralnetwork models on devices with limited memory capacity, refinedmethods for compressing these are essential. This project aimsat investigating some possible solutions, namely pruning andpartitioned logit based knowledge distillation, using teacherstudentlearning methods. A cumbersome benchmark teacherneural network was developed and used as a reference. A specialcase of logit based teacher-student learning was then applied,resulting not only in a compressed model, but also in a convenientway of distributing it. The individual student models were ableto mimic the parts of the teacher model with small losses, whilethe network of student models achieved similar accuracy as theteacher model. Overall, the size of the network of student modelswas around 11% of the teacher. Another popular method ofcompressing neural networks was also tested - pruning. Pruningthe teacher network resulted in a much smaller model, around18% of the teacher model, with similar accuracy. För att möjliggöra användning av storaneurala nätverksmodeller på enheter med begränsad minneskapacitetkrävs raffinerade metoder för komprimering av dessa.Detta projekt syftar till att undersöka några möjliga lösningar,nämligen pruning och partitionerad logit-baserad knowledgedistillation, med hjälp av teacher-student-träning. Ett stortriktmärkesnätverk utvecklades och användes som referens. Enspeciell typ av logit-baserad teacher-student-träning tillämpadessedan, vilket inte bara resulterade i en komprimerad modellutan också i ett smidigt sätt att distribuera den på. De enskildastudent-modellerna kunde efterlikna delar av teachermodellenmed små förluster, medan nätverket av studentmodelleruppnådde ungefär samma noggrannhet som teachermodellen.Sammantaget uppmättes storleken av nätverket avstudent-modeller till cirka 11 % av teacher-modellen. En annanpopulär metod för komprimering av neurala nätverk testadesockså pruning. Pruning av teacher-modellen resulterade i enmycket mindre modell, cirka 18 % av ...
format Bachelor Thesis
author Backe, Hannes
Rydberg, David
author_facet Backe, Hannes
Rydberg, David
author_sort Backe, Hannes
title Compression and Distribution of a Neural Network With IoT Applications
title_short Compression and Distribution of a Neural Network With IoT Applications
title_full Compression and Distribution of a Neural Network With IoT Applications
title_fullStr Compression and Distribution of a Neural Network With IoT Applications
title_full_unstemmed Compression and Distribution of a Neural Network With IoT Applications
title_sort compression and distribution of a neural network with iot applications
publisher KTH, Skolan för elektroteknik och datavetenskap (EECS)
publishDate 2021
url http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-308460
genre DML
genre_facet DML
op_relation TRITA-EECS-EX
2021:186
http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-308460
op_rights info:eu-repo/semantics/openAccess
_version_ 1766397715299696640