pkgsrc-WIP-changes archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
word2vec: Import version 0.1c
Module Name: pkgsrc-wip
Committed By: Min Sik Kim <minskim%NetBSD.org@localhost>
Pushed By: minskim
Date: Wed Jun 6 10:49:05 2018 -0700
Changeset: d1239b26c5e27b63bba82550cadfa31d0642e170
Modified Files:
Makefile
Added Files:
word2vec/DESCR
word2vec/Makefile
word2vec/PLIST
word2vec/distinfo
word2vec/patches/patch-makefile
word2vec/patches/patch-word2phrase.c
word2vec/patches/patch-word2vec.c
Log Message:
word2vec: Import version 0.1c
word2vec is an implementation of the Continuous Bag-of-Words (CBOW)
and the Skip-gram model (SG), as well as several demo scripts. Given
a text corpus, the word2vec tool learns a vector for every word in the
vocabulary using the Continuous Bag-of-Words or the Skip-Gram neural
network architectures.
To see a diff of this commit:
https://wip.pkgsrc.org/cgi-bin/gitweb.cgi?p=pkgsrc-wip.git;a=commitdiff;h=d1239b26c5e27b63bba82550cadfa31d0642e170
Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.
diffstat:
Makefile | 1 +
word2vec/DESCR | 5 +++++
word2vec/Makefile | 22 +++++++++++++++++++
word2vec/PLIST | 6 ++++++
word2vec/distinfo | 9 ++++++++
word2vec/patches/patch-makefile | 15 +++++++++++++
word2vec/patches/patch-word2phrase.c | 41 ++++++++++++++++++++++++++++++++++++
word2vec/patches/patch-word2vec.c | 16 ++++++++++++++
8 files changed, 115 insertions(+)
diffs:
diff --git a/Makefile b/Makefile
index 636ab76ad0..39740ed006 100644
--- a/Makefile
+++ b/Makefile
@@ -4601,6 +4601,7 @@ SUBDIR+= wmswallow
SUBDIR+= wmx
SUBDIR+= wmx-gnome
SUBDIR+= wolf4SDL
+SUBDIR+= word2vec
SUBDIR+= workrave
SUBDIR+= wp2md-git
SUBDIR+= wpro
diff --git a/word2vec/DESCR b/word2vec/DESCR
new file mode 100644
index 0000000000..32dea195e2
--- /dev/null
+++ b/word2vec/DESCR
@@ -0,0 +1,5 @@
+word2vec is an implementation of the Continuous Bag-of-Words (CBOW)
+and the Skip-gram model (SG), as well as several demo scripts. Given
+a text corpus, the word2vec tool learns a vector for every word in the
+vocabulary using the Continuous Bag-of-Words or the Skip-Gram neural
+network architectures.
diff --git a/word2vec/Makefile b/word2vec/Makefile
new file mode 100644
index 0000000000..25b1c98780
--- /dev/null
+++ b/word2vec/Makefile
@@ -0,0 +1,22 @@
+# $NetBSD$
+
+DISTNAME= word2vec-0.1c
+CATEGORIES= textproc
+MASTER_SITES= ${MASTER_SITE_GITHUB:=tmikolov/}
+GITHUB_TAG= 20c129af10659f7c50e86e3be406df663beff438
+
+MAINTAINER= minskim%NetBSD.org@localhost
+HOMEPAGE= https://github.com/tmikolov/word2vec
+COMMENT= Tools for computing distributed representtion of words
+LICENSE= apache-2.0
+
+NO_CONFIGURE= yes
+
+INSTALLATION_DIRS+= bin
+
+do-install:
+.for cmd in compute-accuracy distance word2phrase word2vec word-analogy
+ ${INSTALL_PROGRAM} ${WRKSRC}/${cmd} ${DESTDIR}${PREFIX}/bin
+.endfor
+
+.include "../../mk/bsd.pkg.mk"
diff --git a/word2vec/PLIST b/word2vec/PLIST
new file mode 100644
index 0000000000..4f5e05f3a0
--- /dev/null
+++ b/word2vec/PLIST
@@ -0,0 +1,6 @@
+@comment $NetBSD$
+bin/compute-accuracy
+bin/distance
+bin/word-analogy
+bin/word2phrase
+bin/word2vec
diff --git a/word2vec/distinfo b/word2vec/distinfo
new file mode 100644
index 0000000000..aca417878f
--- /dev/null
+++ b/word2vec/distinfo
@@ -0,0 +1,9 @@
+$NetBSD$
+
+SHA1 (word2vec-0.1c-20c129af10659f7c50e86e3be406df663beff438.tar.gz) = 4f0e872348d60223ba3b8412c0b9ccd7dbd07551
+RMD160 (word2vec-0.1c-20c129af10659f7c50e86e3be406df663beff438.tar.gz) = de98886c52303242566eacd5a3eaf4459026bd71
+SHA512 (word2vec-0.1c-20c129af10659f7c50e86e3be406df663beff438.tar.gz) = 698fa7e2e3ce3be4e4ecbe59bfe7f83640f4bc004b089b2b2cd9daa8233e98fbc5b541433317c647a0c796dd9aa2cd3aa186a1f8287e9f536104ed5fc6c1f65c
+Size (word2vec-0.1c-20c129af10659f7c50e86e3be406df663beff438.tar.gz) = 104875 bytes
+SHA1 (patch-makefile) = 2e32c5af8922008c2961fb2a7a4f59fd31ae0df9
+SHA1 (patch-word2phrase.c) = 47ccf0897b76960a6ef48ddfffc60cc4c59afaee
+SHA1 (patch-word2vec.c) = 1f0e2cf42c6156268f60075aa0a60ab750bc8bfd
diff --git a/word2vec/patches/patch-makefile b/word2vec/patches/patch-makefile
new file mode 100644
index 0000000000..d92716a79c
--- /dev/null
+++ b/word2vec/patches/patch-makefile
@@ -0,0 +1,15 @@
+$NetBSD$
+
+Do not override compiler set by pkgsrc.
+
+--- makefile.orig 2017-07-16 22:46:08.000000000 +0000
++++ makefile
+@@ -1,6 +1,6 @@
+-CC = gcc
++CC?= gcc
+ #Using -Ofast instead of -O3 might result in faster code, but is supported only by newer GCC versions
+-CFLAGS = -lm -pthread -O3 -march=native -Wall -funroll-loops -Wno-unused-result
++CFLAGS+= -lm -pthread -Wall -funroll-loops -Wno-unused-result
+
+ all: word2vec word2phrase distance word-analogy compute-accuracy
+
diff --git a/word2vec/patches/patch-word2phrase.c b/word2vec/patches/patch-word2phrase.c
new file mode 100644
index 0000000000..775f1d129a
--- /dev/null
+++ b/word2vec/patches/patch-word2phrase.c
@@ -0,0 +1,41 @@
+$NetBSD$
+
+Portability fixes.
+https://github.com/tmikolov/word2vec/pull/40
+
+--- word2phrase.c.orig 2017-07-16 22:46:08.000000000 +0000
++++ word2phrase.c
+@@ -42,7 +42,7 @@ unsigned long long next_random = 1;
+ void ReadWord(char *word, FILE *fin, char *eof) {
+ int a = 0, ch;
+ while (1) {
+- ch = fgetc_unlocked(fin);
++ ch = getc_unlocked(fin);
+ if (ch == EOF) {
+ *eof = 1;
+ break;
+@@ -246,7 +246,7 @@ void TrainModel() {
+ if (eof) break;
+ if (word[0] == '\n') {
+ //fprintf(fo, "\n");
+- fputc_unlocked('\n', fo);
++ putc_unlocked('\n', fo);
+ continue;
+ }
+ cn++;
+@@ -286,12 +286,12 @@ void TrainModel() {
+ next_random = next_random * (unsigned long long)25214903917 + 11;
+ //if (next_random & 0x10000) score = 0;
+ if (score > threshold) {
+- fputc_unlocked('_', fo);
++ putc_unlocked('_', fo);
+ pb = 0;
+- } else fputc_unlocked(' ', fo);
++ } else putc_unlocked(' ', fo);
+ a = 0;
+ while (word[a]) {
+- fputc_unlocked(word[a], fo);
++ putc_unlocked(word[a], fo);
+ a++;
+ }
+ pa = pb;
diff --git a/word2vec/patches/patch-word2vec.c b/word2vec/patches/patch-word2vec.c
new file mode 100644
index 0000000000..6b43214d55
--- /dev/null
+++ b/word2vec/patches/patch-word2vec.c
@@ -0,0 +1,16 @@
+$NetBSD$
+
+Portability fix.
+https://github.com/tmikolov/word2vec/pull/40
+
+--- word2vec.c.orig 2017-07-16 22:46:08.000000000 +0000
++++ word2vec.c
+@@ -71,7 +71,7 @@ void InitUnigramTable() {
+ void ReadWord(char *word, FILE *fin, char *eof) {
+ int a = 0, ch;
+ while (1) {
+- ch = fgetc_unlocked(fin);
++ ch = getc_unlocked(fin);
+ if (ch == EOF) {
+ *eof = 1;
+ break;
Home |
Main Index |
Thread Index |
Old Index