國立虎尾科技大學 |

Empowering Large Language Models With Efficient and Automated Systems.

Record Type:	Language materials, manuscript : Monograph/item
Title/Author:	Empowering Large Language Models With Efficient and Automated Systems./
Author:	Li, Zhuohan.
Description:	1 online resource (153 pages)
Notes:	Source: Dissertations Abstracts International, Volume: 86-03, Section: A.
Contained By:	Dissertations Abstracts International86-03A.
Subject:	Computer science. -
Online resource:	click for full text (PQDT)
ISBN:	9798384449218

Empowering Large Language Models With Efficient and Automated Systems.
Li, Zhuohan.

Empowering Large Language Models With Efficient and Automated Systems. - 1 online resource (153 pages)

Source: Dissertations Abstracts International, Volume: 86-03, Section: A.

Thesis (Ph.D.)--University of California, Berkeley, 2024.

Includes bibliographical references

Large Language Models (LLMs) have shown remarkable capabilities in a variety of tasks, including chatting, programming, and searching. However, the high costs of LLMs are preventing these models from being deployed for the vast majority of applications. In this dissertation, we focus on building efficient and automated systems to reduce costs and democratize access to large language models. We first introduce systems to optimize computational efficiency and reduce the engineering overhead for distributed LLM training. We develop TeraPipe, which proposes a new dimension to perform pipeline parallel training for LLMs, and also Alpa, the world's first compiler capable of automatically distributing arbitrary neural networks with all existing parallelization methods. While training is typically a one-time cost, deploying and serving an LLM requires running LLM inference continuously, which is the top blocker for the real-world deployment of LLMs. We improve the serving scalability with AlpaServe through model parallelism, and increase the memory utilization and the LLM inference throughput with a new attention algorithm, PagedAttention, and an end-to-end serving system, vLLM. Overall, these systems provide comprehensive solutions that significantly improve both training and inference efficiency for large language models. Together, these systems lower the high costs associated with large language models, democratizing their deployment across various real-world applications.

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

Mode of access: World Wide Web

ISBN: 9798384449218Subjects--Topical Terms:

573171
Computer science.
Subjects--Index Terms:

Deep learningIndex Terms--Genre/Form:

554714
Electronic books.

Empowering Large Language Models With Efficient and Automated Systems.
LDR:02862ntm a22003977 4500 001 1152770
005 20241213095556.5
006 m o d
007 cr mn ---uuuuu
008 250605s2024 xx obm 000 0 eng d
020 $a 9798384449218
035 $a (MiAaPQ)AAI31296965
035 $a AAI31296965
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Li, Zhuohan. $3 1479856
245 1 0 $a Empowering Large Language Models With Efficient and Automated Systems.
264 0 $c 2024
300 $a 1 online resource (153 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: 86-03, Section: A.
500 $a Advisor: Stoica, Ion.
502 $a Thesis (Ph.D.)--University of California, Berkeley, 2024.
504 $a Includes bibliographical references
520 $a Large Language Models (LLMs) have shown remarkable capabilities in a variety of tasks, including chatting, programming, and searching. However, the high costs of LLMs are preventing these models from being deployed for the vast majority of applications. In this dissertation, we focus on building efficient and automated systems to reduce costs and democratize access to large language models. We first introduce systems to optimize computational efficiency and reduce the engineering overhead for distributed LLM training. We develop TeraPipe, which proposes a new dimension to perform pipeline parallel training for LLMs, and also Alpa, the world's first compiler capable of automatically distributing arbitrary neural networks with all existing parallelization methods. While training is typically a one-time cost, deploying and serving an LLM requires running LLM inference continuously, which is the top blocker for the real-world deployment of LLMs. We improve the serving scalability with AlpaServe through model parallelism, and increase the memory utilization and the LLM inference throughput with a new attention algorithm, PagedAttention, and an end-to-end serving system, vLLM. Overall, these systems provide comprehensive solutions that significantly improve both training and inference efficiency for large language models. Together, these systems lower the high costs associated with large language models, democratizing their deployment across various real-world applications.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2024
538 $a Mode of access: World Wide Web
650 4 $a Computer science. $3 573171
650 4 $a Linguistics. $3 557829
650 4 $a Information technology. $3 559429
653 $a Deep learning
653 $a Distributed systems
653 $a Large language models
653 $a Machine learning
655 7 $a Electronic books. $2 local $3 554714
690 $a 0984
690 $a 0489
690 $a 0800
690 $a 0290
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a University of California, Berkeley. $b Electrical Engineering & Computer Sciences. $3 845564
773 0 $t Dissertations Abstracts International $g 86-03A.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31296965 $z click for full text (PQDT)