-
Notifications
You must be signed in to change notification settings - Fork 2
PengjieRen/LibSum
Folders and files
Name | Name | Last commit message | Last commit date | |
---|---|---|---|---|
Repository files navigation
# LibSum 1. Decompression. 1) Decompress /SRSum.7z to /SRSum 2) Decompress /root/ROUGE/RELEASE-1.5.5.7z to /root/ROUGE/RELEASE-1.5.5 3) Decompress /root/model/glove.6B.50d.txt.7z to /root/model/glove.6B.50d.txt 2. ROUGE configuration. 1) Install ActivePerl-5.20.2.2002-MSWin32-x86-64int-299195.msi to C:\Perl\ and make sure C:\Perl\bin\perl.exe exists 2) Install perl package XML-DOM using cpan with the command: cpan XML::DOM 3. Arguments. Usage: LibSum.exe [OPTIONS] [trainFilePath] testFilePath Options: -l, --load=VALUE Default NULL. Saved model file path (e.g., ./root/model/SRSum.2004.model). -q, --qsr=VALUE Default false. Whether with QSRSum. (for DUC 2005, 2006, and 2007). -c, --context=VALUE Default 4. (The number of contexts in SRSum. Increasing this will get better results but will also slow down the speed). -i, --iteration=VALUE Default 50. -t, --threshold=VALUE Default 0.5. -b, --byte=VALUE Default 665 (DUC 2004 ROUGE settings). -w, --word=VALUE (DUC 2001, 2002 ROUGE settings). -h, --help Show this message and exit. 4. Test existing model. For example, test DUC 2004 model: LibSum.exe -l PATH\root\model\SRSum\SRSum.2004.model PATH\root\data\duc2004.test.txt The console outputs: Intel MKL (x64; revision 11; MKL 2017.0) SLF4J: Failed to load class "org.slf4j.impl.StaticLoggerBinder". SLF4J: Defaulting to no-operation (NOP) logger implementation SLF4J: See http://www.slf4j.org/codes.html#StaticLoggerBinder for further details. Reading POS tagger model from edu/stanford/nlp/models/pos-tagger/english-left3words/english-left3words-distsim.tagger ... done [1.0 sec]. [Read] xxxxxxxxxxxxxxx\root\data\duc2004.test.txt 13300 Corpus Count: 50 .................................................. C:\Perl\bin\perl.exe ..\root\Rouge\RELEASE-1.5.5\ROUGE-1.5.5.pl -b 665 -e ..\root\Rouge\RELEASE-1.5.5\data -n 2 -m -2 4 -u -c 95 -x -r 1000 -f A -p 0.5 -t 0 -a -d "..\root\model\SRSum\rouge-summary__test\rouge_config.xml" ROURGE_1_R 0.39285 ROURGE_1_P 0.38792 ROURGE_1_F 0.39017 ROURGE_2_R 0.10698 ROURGE_2_P 0.10572 ROURGE_2_F 0.1063 ROURGE_SU4_R 0.14823 ROURGE_SU4_P 0.14639 ROURGE_SU4_F 0.14723 5. Train new model. For example, train DUC 2004 model: LibSum.exe -q false -c 4 -i 100 -t 0.5 -b 665 PATH\root\data\duc2004.train.txt PATH\root\data\duc2004.test.txt For example, train DUC 2007 model: LibSum.exe -q true -c 4 -i 100 -t 0.5 -w 250 PATH\root\data\duc2007.train.txt PATH\root\data\duc2007.test.txt
About
Sentence Realation based Summarization (SRSum). It's an improved version of Contextual Relation based Summarization (CRSum)
Resources
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published