CharPoet: A Chinese Classical Poetry Generation System Based on Token-free LLM

Automatic Chinese classical poetry generation has attracted much research interest, but achieving effective control over format and content simultaneously remains challenging. Traditional systems usually accept keywords as user inputs, resulting in limited control over content. Large language models...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Yu, Chengyue, Zang, Lei, Wang, Jiaotuan, Zhuang, Chenyi, Gu, Jinjie
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Yu, Chengyue
Zang, Lei
Wang, Jiaotuan
Zhuang, Chenyi
Gu, Jinjie
description Automatic Chinese classical poetry generation has attracted much research interest, but achieving effective control over format and content simultaneously remains challenging. Traditional systems usually accept keywords as user inputs, resulting in limited control over content. Large language models (LLMs) improve content control by allowing unrestricted user instructions, but the token-by-token generation process frequently makes format errors. Motivated by this, we propose CharPoet, a Chinese classical poetry generation system based on token-free LLM, which provides effective control over both format and content. Our token-free architecture generates in a character-by-character manner, enabling precise control over the number of characters. Pruned from existing token-based LLMs, CharPoet inherits their pretrained capabilities and can generate poetry following instructions like "Write me a poem for my mother's birthday." CharPoet achieves format accuracy above 0.96, outperforming Jiuge-GPT-2 (0.91) and GPT-4 (0.38). In terms of content quality, CharPoet surpasses traditional systems including Jiuge, and is comparable to other LLMs. Our system is open source and available at https://modelscope.cn/models/CharPoet/CharPoet. A video demonstration of CharPoet is available at https://youtu.be/voZ25qEp3Dc.
doi_str_mv 10.48550/arxiv.2401.03512
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2401_03512</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2401_03512</sourcerecordid><originalsourceid>FETCH-arxiv_primary_2401_035123</originalsourceid><addsrcrecordid>eNpjYJA0NNAzsTA1NdBPLKrILNMzMjEw1DMwNjU04mTwc85ILArITy2xUnBUcM7IzEstTlVwzkksLs5MTsxRAMkUVSq4p-alFiWWZObnKQRXFpek5io4JRanpigA-SH52al5umlFqakKPj6-PAysaYk5xam8UJqbQd7NNcTZQxdsc3xBUWZuYlFlPMgF8WAXGBNWAQDJDDpc</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>CharPoet: A Chinese Classical Poetry Generation System Based on Token-free LLM</title><source>arXiv.org</source><creator>Yu, Chengyue ; Zang, Lei ; Wang, Jiaotuan ; Zhuang, Chenyi ; Gu, Jinjie</creator><creatorcontrib>Yu, Chengyue ; Zang, Lei ; Wang, Jiaotuan ; Zhuang, Chenyi ; Gu, Jinjie</creatorcontrib><description>Automatic Chinese classical poetry generation has attracted much research interest, but achieving effective control over format and content simultaneously remains challenging. Traditional systems usually accept keywords as user inputs, resulting in limited control over content. Large language models (LLMs) improve content control by allowing unrestricted user instructions, but the token-by-token generation process frequently makes format errors. Motivated by this, we propose CharPoet, a Chinese classical poetry generation system based on token-free LLM, which provides effective control over both format and content. Our token-free architecture generates in a character-by-character manner, enabling precise control over the number of characters. Pruned from existing token-based LLMs, CharPoet inherits their pretrained capabilities and can generate poetry following instructions like "Write me a poem for my mother's birthday." CharPoet achieves format accuracy above 0.96, outperforming Jiuge-GPT-2 (0.91) and GPT-4 (0.38). In terms of content quality, CharPoet surpasses traditional systems including Jiuge, and is comparable to other LLMs. Our system is open source and available at https://modelscope.cn/models/CharPoet/CharPoet. A video demonstration of CharPoet is available at https://youtu.be/voZ25qEp3Dc.</description><identifier>DOI: 10.48550/arxiv.2401.03512</identifier><language>eng</language><subject>Computer Science - Artificial Intelligence ; Computer Science - Computation and Language ; Computer Science - Learning</subject><creationdate>2024-01</creationdate><rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,777,882</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2401.03512$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2401.03512$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Yu, Chengyue</creatorcontrib><creatorcontrib>Zang, Lei</creatorcontrib><creatorcontrib>Wang, Jiaotuan</creatorcontrib><creatorcontrib>Zhuang, Chenyi</creatorcontrib><creatorcontrib>Gu, Jinjie</creatorcontrib><title>CharPoet: A Chinese Classical Poetry Generation System Based on Token-free LLM</title><description>Automatic Chinese classical poetry generation has attracted much research interest, but achieving effective control over format and content simultaneously remains challenging. Traditional systems usually accept keywords as user inputs, resulting in limited control over content. Large language models (LLMs) improve content control by allowing unrestricted user instructions, but the token-by-token generation process frequently makes format errors. Motivated by this, we propose CharPoet, a Chinese classical poetry generation system based on token-free LLM, which provides effective control over both format and content. Our token-free architecture generates in a character-by-character manner, enabling precise control over the number of characters. Pruned from existing token-based LLMs, CharPoet inherits their pretrained capabilities and can generate poetry following instructions like "Write me a poem for my mother's birthday." CharPoet achieves format accuracy above 0.96, outperforming Jiuge-GPT-2 (0.91) and GPT-4 (0.38). In terms of content quality, CharPoet surpasses traditional systems including Jiuge, and is comparable to other LLMs. Our system is open source and available at https://modelscope.cn/models/CharPoet/CharPoet. A video demonstration of CharPoet is available at https://youtu.be/voZ25qEp3Dc.</description><subject>Computer Science - Artificial Intelligence</subject><subject>Computer Science - Computation and Language</subject><subject>Computer Science - Learning</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNpjYJA0NNAzsTA1NdBPLKrILNMzMjEw1DMwNjU04mTwc85ILArITy2xUnBUcM7IzEstTlVwzkksLs5MTsxRAMkUVSq4p-alFiWWZObnKQRXFpek5io4JRanpigA-SH52al5umlFqakKPj6-PAysaYk5xam8UJqbQd7NNcTZQxdsc3xBUWZuYlFlPMgF8WAXGBNWAQDJDDpc</recordid><startdate>20240107</startdate><enddate>20240107</enddate><creator>Yu, Chengyue</creator><creator>Zang, Lei</creator><creator>Wang, Jiaotuan</creator><creator>Zhuang, Chenyi</creator><creator>Gu, Jinjie</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20240107</creationdate><title>CharPoet: A Chinese Classical Poetry Generation System Based on Token-free LLM</title><author>Yu, Chengyue ; Zang, Lei ; Wang, Jiaotuan ; Zhuang, Chenyi ; Gu, Jinjie</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-arxiv_primary_2401_035123</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Computer Science - Artificial Intelligence</topic><topic>Computer Science - Computation and Language</topic><topic>Computer Science - Learning</topic><toplevel>online_resources</toplevel><creatorcontrib>Yu, Chengyue</creatorcontrib><creatorcontrib>Zang, Lei</creatorcontrib><creatorcontrib>Wang, Jiaotuan</creatorcontrib><creatorcontrib>Zhuang, Chenyi</creatorcontrib><creatorcontrib>Gu, Jinjie</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Yu, Chengyue</au><au>Zang, Lei</au><au>Wang, Jiaotuan</au><au>Zhuang, Chenyi</au><au>Gu, Jinjie</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>CharPoet: A Chinese Classical Poetry Generation System Based on Token-free LLM</atitle><date>2024-01-07</date><risdate>2024</risdate><abstract>Automatic Chinese classical poetry generation has attracted much research interest, but achieving effective control over format and content simultaneously remains challenging. Traditional systems usually accept keywords as user inputs, resulting in limited control over content. Large language models (LLMs) improve content control by allowing unrestricted user instructions, but the token-by-token generation process frequently makes format errors. Motivated by this, we propose CharPoet, a Chinese classical poetry generation system based on token-free LLM, which provides effective control over both format and content. Our token-free architecture generates in a character-by-character manner, enabling precise control over the number of characters. Pruned from existing token-based LLMs, CharPoet inherits their pretrained capabilities and can generate poetry following instructions like "Write me a poem for my mother's birthday." CharPoet achieves format accuracy above 0.96, outperforming Jiuge-GPT-2 (0.91) and GPT-4 (0.38). In terms of content quality, CharPoet surpasses traditional systems including Jiuge, and is comparable to other LLMs. Our system is open source and available at https://modelscope.cn/models/CharPoet/CharPoet. A video demonstration of CharPoet is available at https://youtu.be/voZ25qEp3Dc.</abstract><doi>10.48550/arxiv.2401.03512</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2401.03512
ispartof
issn
language eng
recordid cdi_arxiv_primary_2401_03512
source arXiv.org
subjects Computer Science - Artificial Intelligence
Computer Science - Computation and Language
Computer Science - Learning
title CharPoet: A Chinese Classical Poetry Generation System Based on Token-free LLM
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-19T05%3A17%3A09IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=CharPoet:%20A%20Chinese%20Classical%20Poetry%20Generation%20System%20Based%20on%20Token-free%20LLM&rft.au=Yu,%20Chengyue&rft.date=2024-01-07&rft_id=info:doi/10.48550/arxiv.2401.03512&rft_dat=%3Carxiv_GOX%3E2401_03512%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true