diff options
author | Marcin Junczys-Dowmunt <junczys@amu.edu.pl> | 2014-12-19 22:31:23 +0300 |
---|---|---|
committer | Marcin Junczys-Dowmunt <junczys@amu.edu.pl> | 2014-12-19 22:31:23 +0300 |
commit | 5d85f203dc84b5b516379ceebac40ec0c5d0f5bc (patch) | |
tree | cabe06b2e02ac73c3da5945300f959eab66c8fff | |
parent | 84fa79684b5216b0d7a276cb09aff4fc5eecce94 (diff) |
build word2vec by default
-rw-r--r-- | Jamroot | 1 | ||||
-rw-r--r-- | contrib/word2vec/Jamfile | 4 | ||||
-rw-r--r-- | contrib/word2vec/README.txt | 21 | ||||
-rw-r--r-- | contrib/word2vec/compute-accuracy.c | 137 | ||||
-rwxr-xr-x | contrib/word2vec/demo-analogy.sh | 11 | ||||
-rwxr-xr-x | contrib/word2vec/demo-classes.sh | 8 | ||||
-rwxr-xr-x | contrib/word2vec/demo-phrase-accuracy.sh | 12 | ||||
-rwxr-xr-x | contrib/word2vec/demo-phrases.sh | 8 | ||||
-rwxr-xr-x | contrib/word2vec/demo-word-accuracy.sh | 8 | ||||
-rwxr-xr-x | contrib/word2vec/demo-word.sh | 7 | ||||
-rw-r--r-- | contrib/word2vec/distance.c | 136 | ||||
-rw-r--r-- | contrib/word2vec/makefile | 20 | ||||
-rw-r--r-- | contrib/word2vec/word-analogy.c | 138 | ||||
-rw-r--r-- | contrib/word2vec/word2phrase.c | 292 |
14 files changed, 5 insertions, 798 deletions
@@ -212,6 +212,7 @@ biconcor mira//mira contrib/server//mosesserver contrib/bleu-champ//programs +contrib/word2vec//word2vec mm ; diff --git a/contrib/word2vec/Jamfile b/contrib/word2vec/Jamfile new file mode 100644 index 000000000..d3c3efd76 --- /dev/null +++ b/contrib/word2vec/Jamfile @@ -0,0 +1,4 @@ + +exe word2vec : word2vec.c ; + +install legacy : word2vec : <location>. ; diff --git a/contrib/word2vec/README.txt b/contrib/word2vec/README.txt deleted file mode 100644 index 2bcd9da44..000000000 --- a/contrib/word2vec/README.txt +++ /dev/null @@ -1,21 +0,0 @@ -Tools for computing distributed representtion of words ------------------------------------------------------- - -We provide 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. The user should to specify the following: - - desired vector dimensionality - - the size of the context window for either the Skip-Gram or the Continuous Bag-of-Words model - - training algorithm: hierarchical softmax and / or negative sampling - - threshold for downsampling the frequent words - - number of threads to use - - the format of the output word vector file (text or binary) - -Usually, the other hyper-parameters such as the learning rate do not need to be tuned for different training sets. - -The script demo-word.sh downloads a small (100MB) text corpus from the web, and trains a small word vector model. After the training -is finished, the user can interactively explore the similarity of the words. - -More information about the scripts is provided at https://code.google.com/p/word2vec/ - diff --git a/contrib/word2vec/compute-accuracy.c b/contrib/word2vec/compute-accuracy.c deleted file mode 100644 index d83fcbb9a..000000000 --- a/contrib/word2vec/compute-accuracy.c +++ /dev/null @@ -1,137 +0,0 @@ -// Copyright 2013 Google Inc. All Rights Reserved. -// -// Licensed under the Apache License, Version 2.0 (the "License"); -// you may not use this file except in compliance with the License. -// You may obtain a copy of the License at -// -// http://www.apache.org/licenses/LICENSE-2.0 -// -// Unless required by applicable law or agreed to in writing, software -// distributed under the License is distributed on an "AS IS" BASIS, -// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -// See the License for the specific language governing permissions and -// limitations under the License. - -#include <stdio.h> -#include <stdlib.h> -#include <string.h> -#include <math.h> -#include <malloc.h> -#include <ctype.h> - -const long long max_size = 2000; // max length of strings -const long long N = 1; // number of closest words -const long long max_w = 50; // max length of vocabulary entries - -int main(int argc, char **argv) -{ - FILE *f; - char st1[max_size], st2[max_size], st3[max_size], st4[max_size], bestw[N][max_size], file_name[max_size], ch; - float dist, len, bestd[N], vec[max_size]; - long long words, size, a, b, c, d, b1, b2, b3, threshold = 0; - float *M; - char *vocab; - int TCN, CCN = 0, TACN = 0, CACN = 0, SECN = 0, SYCN = 0, SEAC = 0, SYAC = 0, QID = 0, TQ = 0, TQS = 0; - if (argc < 2) { - printf("Usage: ./compute-accuracy <FILE> <threshold>\nwhere FILE contains word projections, and threshold is used to reduce vocabulary of the model for fast approximate evaluation (0 = off, otherwise typical value is 30000)\n"); - return 0; - } - strcpy(file_name, argv[1]); - if (argc > 2) threshold = atoi(argv[2]); - f = fopen(file_name, "rb"); - if (f == NULL) { - printf("Input file not found\n"); - return -1; - } - fscanf(f, "%lld", &words); - if (threshold) if (words > threshold) words = threshold; - fscanf(f, "%lld", &size); - vocab = (char *)malloc(words * max_w * sizeof(char)); - M = (float *)malloc(words * size * sizeof(float)); - if (M == NULL) { - printf("Cannot allocate memory: %lld MB\n", words * size * sizeof(float) / 1048576); - return -1; - } - for (b = 0; b < words; b++) { - fscanf(f, "%s%c", &vocab[b * max_w], &ch); - for (a = 0; a < max_w; a++) vocab[b * max_w + a] = toupper(vocab[b * max_w + a]); - for (a = 0; a < size; a++) fread(&M[a + b * size], sizeof(float), 1, f); - len = 0; - for (a = 0; a < size; a++) len += M[a + b * size] * M[a + b * size]; - len = sqrt(len); - for (a = 0; a < size; a++) M[a + b * size] /= len; - } - fclose(f); - TCN = 0; - while (1) { - for (a = 0; a < N; a++) bestd[a] = 0; - for (a = 0; a < N; a++) bestw[a][0] = 0; - scanf("%s", st1); - for (a = 0; a < strlen(st1); a++) st1[a] = toupper(st1[a]); - if ((!strcmp(st1, ":")) || (!strcmp(st1, "EXIT")) || feof(stdin)) { - if (TCN == 0) TCN = 1; - if (QID != 0) { - printf("ACCURACY TOP1: %.2f %% (%d / %d)\n", CCN / (float)TCN * 100, CCN, TCN); - printf("Total accuracy: %.2f %% Semantic accuracy: %.2f %% Syntactic accuracy: %.2f %% \n", CACN / (float)TACN * 100, SEAC / (float)SECN * 100, SYAC / (float)SYCN * 100); - } - QID++; - scanf("%s", st1); - if (feof(stdin)) break; - printf("%s:\n", st1); - TCN = 0; - CCN = 0; - continue; - } - if (!strcmp(st1, "EXIT")) break; - scanf("%s", st2); - for (a = 0; a < strlen(st2); a++) st2[a] = toupper(st2[a]); - scanf("%s", st3); - for (a = 0; a<strlen(st3); a++) st3[a] = toupper(st3[a]); - scanf("%s", st4); - for (a = 0; a < strlen(st4); a++) st4[a] = toupper(st4[a]); - for (b = 0; b < words; b++) if (!strcmp(&vocab[b * max_w], st1)) break; - b1 = b; - for (b = 0; b < words; b++) if (!strcmp(&vocab[b * max_w], st2)) break; - b2 = b; - for (b = 0; b < words; b++) if (!strcmp(&vocab[b * max_w], st3)) break; - b3 = b; - for (a = 0; a < N; a++) bestd[a] = 0; - for (a = 0; a < N; a++) bestw[a][0] = 0; - TQ++; - if (b1 == words) continue; - if (b2 == words) continue; - if (b3 == words) continue; - for (b = 0; b < words; b++) if (!strcmp(&vocab[b * max_w], st4)) break; - if (b == words) continue; - for (a = 0; a < size; a++) vec[a] = (M[a + b2 * size] - M[a + b1 * size]) + M[a + b3 * size]; - TQS++; - for (c = 0; c < words; c++) { - if (c == b1) continue; - if (c == b2) continue; - if (c == b3) continue; - dist = 0; - for (a = 0; a < size; a++) dist += vec[a] * M[a + c * size]; - for (a = 0; a < N; a++) { - if (dist > bestd[a]) { - for (d = N - 1; d > a; d--) { - bestd[d] = bestd[d - 1]; - strcpy(bestw[d], bestw[d - 1]); - } - bestd[a] = dist; - strcpy(bestw[a], &vocab[c * max_w]); - break; - } - } - } - if (!strcmp(st4, bestw[0])) { - CCN++; - CACN++; - if (QID <= 5) SEAC++; else SYAC++; - } - if (QID <= 5) SECN++; else SYCN++; - TCN++; - TACN++; - } - printf("Questions seen / total: %d %d %.2f %% \n", TQS, TQ, TQS/(float)TQ*100); - return 0; -} diff --git a/contrib/word2vec/demo-analogy.sh b/contrib/word2vec/demo-analogy.sh deleted file mode 100755 index 3d1fcf7f9..000000000 --- a/contrib/word2vec/demo-analogy.sh +++ /dev/null @@ -1,11 +0,0 @@ -make -if [ ! -e text8 ]; then - wget http://mattmahoney.net/dc/text8.zip -O text8.gz - gzip -d text8.gz -f -fi -echo ----------------------------------------------------------------------------------------------------- -echo Note that for the word analogy to perform well, the models should be trained on much larger data sets -echo Example input: paris france berlin -echo ----------------------------------------------------------------------------------------------------- -time ./word2vec -train text8 -output vectors.bin -cbow 0 -size 200 -window 5 -negative 0 -hs 1 -sample 1e-3 -threads 12 -binary 1 -./word-analogy vectors.bin diff --git a/contrib/word2vec/demo-classes.sh b/contrib/word2vec/demo-classes.sh deleted file mode 100755 index b0b9d991c..000000000 --- a/contrib/word2vec/demo-classes.sh +++ /dev/null @@ -1,8 +0,0 @@ -make -if [ ! -e text8 ]; then - wget http://mattmahoney.net/dc/text8.zip -O text8.gz - gzip -d text8.gz -f -fi -time ./word2vec -train text8 -output classes.txt -cbow 0 -size 200 -window 5 -negative 0 -hs 1 -sample 1e-3 -threads 12 -classes 500 -sort classes.txt -k 2 -n > classes.sorted.txt -echo The word classes were saved to file classes.sorted.txt diff --git a/contrib/word2vec/demo-phrase-accuracy.sh b/contrib/word2vec/demo-phrase-accuracy.sh deleted file mode 100755 index eb2b392e8..000000000 --- a/contrib/word2vec/demo-phrase-accuracy.sh +++ /dev/null @@ -1,12 +0,0 @@ -make -if [ ! -e text8 ]; then - wget http://mattmahoney.net/dc/text8.zip -O text8.gz - gzip -d text8.gz -f -fi -echo ---------------------------------------------------------------------------------------------------------------- -echo Note that the accuracy and coverage of the test set questions is going to be low with this small training corpus -echo To achieve better accuracy, larger training set is needed -echo ---------------------------------------------------------------------------------------------------------------- -time ./word2phrase -train text8 -output text8-phrase -threshold 500 -debug 2 -min-count 3 -time ./word2vec -train text8-phrase -output vectors-phrase.bin -cbow 0 -size 300 -window 10 -negative 0 -hs 1 -sample 1e-3 -threads 12 -binary 1 -min-count 3 -./compute-accuracy vectors-phrase.bin <questions-phrases.txt diff --git a/contrib/word2vec/demo-phrases.sh b/contrib/word2vec/demo-phrases.sh deleted file mode 100755 index c833b8149..000000000 --- a/contrib/word2vec/demo-phrases.sh +++ /dev/null @@ -1,8 +0,0 @@ -make -if [ ! -e text8 ]; then - wget http://mattmahoney.net/dc/text8.zip -O text8.gz - gzip -d text8.gz -f -fi -time ./word2phrase -train text8 -output text8-phrase -threshold 500 -debug 2 -time ./word2vec -train text8-phrase -output vectors-phrase.bin -cbow 0 -size 300 -window 10 -negative 0 -hs 1 -sample 1e-3 -threads 12 -binary 1 -./distance vectors-phrase.bin
\ No newline at end of file diff --git a/contrib/word2vec/demo-word-accuracy.sh b/contrib/word2vec/demo-word-accuracy.sh deleted file mode 100755 index ffe828ab1..000000000 --- a/contrib/word2vec/demo-word-accuracy.sh +++ /dev/null @@ -1,8 +0,0 @@ -make -if [ ! -e text8 ]; then - wget http://mattmahoney.net/dc/text8.zip -O text8.gz - gzip -d text8.gz -f -fi -time ./word2vec -train text8 -output vectors.bin -cbow 0 -size 200 -window 5 -negative 0 -hs 1 -sample 1e-3 -threads 12 -binary 1 -./compute-accuracy vectors.bin 30000 < questions-words.txt -# to compute accuracy with the full vocabulary, use: ./compute-accuracy vectors.bin < questions-words.txt diff --git a/contrib/word2vec/demo-word.sh b/contrib/word2vec/demo-word.sh deleted file mode 100755 index 0df5bd510..000000000 --- a/contrib/word2vec/demo-word.sh +++ /dev/null @@ -1,7 +0,0 @@ -make -if [ ! -e text8 ]; then - wget http://mattmahoney.net/dc/text8.zip -O text8.gz - gzip -d text8.gz -f -fi -time ./word2vec -train text8 -output vectors.bin -cbow 0 -size 200 -window 5 -negative 0 -hs 1 -sample 1e-3 -threads 12 -binary 1 -./distance vectors.bin
\ No newline at end of file diff --git a/contrib/word2vec/distance.c b/contrib/word2vec/distance.c deleted file mode 100644 index fbeb24a66..000000000 --- a/contrib/word2vec/distance.c +++ /dev/null @@ -1,136 +0,0 @@ -// Copyright 2013 Google Inc. All Rights Reserved. -// -// Licensed under the Apache License, Version 2.0 (the "License"); -// you may not use this file except in compliance with the License. -// You may obtain a copy of the License at -// -// http://www.apache.org/licenses/LICENSE-2.0 -// -// Unless required by applicable law or agreed to in writing, software -// distributed under the License is distributed on an "AS IS" BASIS, -// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -// See the License for the specific language governing permissions and -// limitations under the License. - -#include <stdio.h> -#include <string.h> -#include <math.h> -#include <malloc.h> - -const long long max_size = 2000; // max length of strings -const long long N = 40; // number of closest words that will be shown -const long long max_w = 50; // max length of vocabulary entries - -int main(int argc, char **argv) { - FILE *f; - char st1[max_size]; - char *bestw[N]; - char file_name[max_size], st[100][max_size]; - float dist, len, bestd[N], vec[max_size]; - long long words, size, a, b, c, d, cn, bi[100]; - char ch; - float *M; - char *vocab; - if (argc < 2) { - printf("Usage: ./distance <FILE>\nwhere FILE contains word projections in the BINARY FORMAT\n"); - return 0; - } - strcpy(file_name, argv[1]); - f = fopen(file_name, "rb"); - if (f == NULL) { - printf("Input file not found\n"); - return -1; - } - fscanf(f, "%lld", &words); - fscanf(f, "%lld", &size); - vocab = (char *)malloc((long long)words * max_w * sizeof(char)); - for (a = 0; a < N; a++) bestw[a] = (char *)malloc(max_size * sizeof(char)); - M = (float *)malloc((long long)words * (long long)size * sizeof(float)); - if (M == NULL) { - printf("Cannot allocate memory: %lld MB %lld %lld\n", (long long)words * size * sizeof(float) / 1048576, words, size); - return -1; - } - for (b = 0; b < words; b++) { - fscanf(f, "%s%c", &vocab[b * max_w], &ch); - for (a = 0; a < size; a++) fread(&M[a + b * size], sizeof(float), 1, f); - len = 0; - for (a = 0; a < size; a++) len += M[a + b * size] * M[a + b * size]; - len = sqrt(len); - for (a = 0; a < size; a++) M[a + b * size] /= len; - } - fclose(f); - while (1) { - for (a = 0; a < N; a++) bestd[a] = 0; - for (a = 0; a < N; a++) bestw[a][0] = 0; - printf("Enter word or sentence (EXIT to break): "); - a = 0; - while (1) { - st1[a] = fgetc(stdin); - if ((st1[a] == '\n') || (a >= max_size - 1)) { - st1[a] = 0; - break; - } - a++; - } - if (!strcmp(st1, "EXIT")) break; - cn = 0; - b = 0; - c = 0; - while (1) { - st[cn][b] = st1[c]; - b++; - c++; - st[cn][b] = 0; - if (st1[c] == 0) break; - if (st1[c] == ' ') { - cn++; - b = 0; - c++; - } - } - cn++; - for (a = 0; a < cn; a++) { - for (b = 0; b < words; b++) if (!strcmp(&vocab[b * max_w], st[a])) break; - if (b == words) b = -1; - bi[a] = b; - printf("\nWord: %s Position in vocabulary: %lld\n", st[a], bi[a]); - if (b == -1) { - printf("Out of dictionary word!\n"); - break; - } - } - if (b == -1) continue; - printf("\n Word Cosine distance\n------------------------------------------------------------------------\n"); - for (a = 0; a < size; a++) vec[a] = 0; - for (b = 0; b < cn; b++) { - if (bi[b] == -1) continue; - for (a = 0; a < size; a++) vec[a] += M[a + bi[b] * size]; - } - len = 0; - for (a = 0; a < size; a++) len += vec[a] * vec[a]; - len = sqrt(len); - for (a = 0; a < size; a++) vec[a] /= len; - for (a = 0; a < N; a++) bestd[a] = -1; - for (a = 0; a < N; a++) bestw[a][0] = 0; - for (c = 0; c < words; c++) { - a = 0; - for (b = 0; b < cn; b++) if (bi[b] == c) a = 1; - if (a == 1) continue; - dist = 0; - for (a = 0; a < size; a++) dist += vec[a] * M[a + c * size]; - for (a = 0; a < N; a++) { - if (dist > bestd[a]) { - for (d = N - 1; d > a; d--) { - bestd[d] = bestd[d - 1]; - strcpy(bestw[d], bestw[d - 1]); - } - bestd[a] = dist; - strcpy(bestw[a], &vocab[c * max_w]); - break; - } - } - } - for (a = 0; a < N; a++) printf("%50s\t\t%f\n", bestw[a], bestd[a]); - } - return 0; -} diff --git a/contrib/word2vec/makefile b/contrib/word2vec/makefile deleted file mode 100644 index b50ee0c3d..000000000 --- a/contrib/word2vec/makefile +++ /dev/null @@ -1,20 +0,0 @@ -CC = gcc -#The -Ofast might not work with older versions of gcc; in that case, use -O2 -CFLAGS = -lm -pthread -Ofast -march=native -Wall -funroll-loops -Wno-unused-result -g - -all: word2vec word2phrase distance word-analogy compute-accuracy - -word2vec : word2vec.c - $(CC) word2vec.c -o word2vec $(CFLAGS) -word2phrase : word2phrase.c - $(CC) word2phrase.c -o word2phrase $(CFLAGS) -distance : distance.c - $(CC) distance.c -o distance $(CFLAGS) -word-analogy : word-analogy.c - $(CC) word-analogy.c -o word-analogy $(CFLAGS) -compute-accuracy : compute-accuracy.c - $(CC) compute-accuracy.c -o compute-accuracy $(CFLAGS) - chmod +x *.sh - -clean: - rm -rf word2vec word2phrase distance word-analogy compute-accuracy
\ No newline at end of file diff --git a/contrib/word2vec/word-analogy.c b/contrib/word2vec/word-analogy.c deleted file mode 100644 index ea91ab141..000000000 --- a/contrib/word2vec/word-analogy.c +++ /dev/null @@ -1,138 +0,0 @@ -// Copyright 2013 Google Inc. All Rights Reserved. -// -// Licensed under the Apache License, Version 2.0 (the "License"); -// you may not use this file except in compliance with the License. -// You may obtain a copy of the License at -// -// http://www.apache.org/licenses/LICENSE-2.0 -// -// Unless required by applicable law or agreed to in writing, software -// distributed under the License is distributed on an "AS IS" BASIS, -// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -// See the License for the specific language governing permissions and -// limitations under the License. - -#include <stdio.h> -#include <string.h> -#include <math.h> -#include <malloc.h> - -const long long max_size = 2000; // max length of strings -const long long N = 40; // number of closest words that will be shown -const long long max_w = 50; // max length of vocabulary entries - -int main(int argc, char **argv) { - FILE *f; - char st1[max_size]; - char bestw[N][max_size]; - char file_name[max_size], st[100][max_size]; - float dist, len, bestd[N], vec[max_size]; - long long words, size, a, b, c, d, cn, bi[100]; - char ch; - float *M; - char *vocab; - if (argc < 2) { - printf("Usage: ./word-analogy <FILE>\nwhere FILE contains word projections in the BINARY FORMAT\n"); - return 0; - } - strcpy(file_name, argv[1]); - f = fopen(file_name, "rb"); - if (f == NULL) { - printf("Input file not found\n"); - return -1; - } - fscanf(f, "%lld", &words); - fscanf(f, "%lld", &size); - vocab = (char *)malloc((long long)words * max_w * sizeof(char)); - M = (float *)malloc((long long)words * (long long)size * sizeof(float)); - if (M == NULL) { - printf("Cannot allocate memory: %lld MB %lld %lld\n", (long long)words * size * sizeof(float) / 1048576, words, size); - return -1; - } - for (b = 0; b < words; b++) { - fscanf(f, "%s%c", &vocab[b * max_w], &ch); - for (a = 0; a < size; a++) fread(&M[a + b * size], sizeof(float), 1, f); - len = 0; - for (a = 0; a < size; a++) len += M[a + b * size] * M[a + b * size]; - len = sqrt(len); - for (a = 0; a < size; a++) M[a + b * size] /= len; - } - fclose(f); - while (1) { - for (a = 0; a < N; a++) bestd[a] = 0; - for (a = 0; a < N; a++) bestw[a][0] = 0; - printf("Enter three words (EXIT to break): "); - a = 0; - while (1) { - st1[a] = fgetc(stdin); - if ((st1[a] == '\n') || (a >= max_size - 1)) { - st1[a] = 0; - break; - } - a++; - } - if (!strcmp(st1, "EXIT")) break; - cn = 0; - b = 0; - c = 0; - while (1) { - st[cn][b] = st1[c]; - b++; - c++; - st[cn][b] = 0; - if (st1[c] == 0) break; - if (st1[c] == ' ') { - cn++; - b = 0; - c++; - } - } - cn++; - if (cn < 3) { - printf("Only %lld words were entered.. three words are needed at the input to perform the calculation\n", cn); - continue; - } - for (a = 0; a < cn; a++) { - for (b = 0; b < words; b++) if (!strcmp(&vocab[b * max_w], st[a])) break; - if (b == words) b = 0; - bi[a] = b; - printf("\nWord: %s Position in vocabulary: %lld\n", st[a], bi[a]); - if (b == 0) { - printf("Out of dictionary word!\n"); - break; - } - } - if (b == 0) continue; - printf("\n Word Distance\n------------------------------------------------------------------------\n"); - for (a = 0; a < size; a++) vec[a] = M[a + bi[1] * size] - M[a + bi[0] * size] + M[a + bi[2] * size]; - len = 0; - for (a = 0; a < size; a++) len += vec[a] * vec[a]; - len = sqrt(len); - for (a = 0; a < size; a++) vec[a] /= len; - for (a = 0; a < N; a++) bestd[a] = 0; - for (a = 0; a < N; a++) bestw[a][0] = 0; - for (c = 0; c < words; c++) { - if (c == bi[0]) continue; - if (c == bi[1]) continue; - if (c == bi[2]) continue; - a = 0; - for (b = 0; b < cn; b++) if (bi[b] == c) a = 1; - if (a == 1) continue; - dist = 0; - for (a = 0; a < size; a++) dist += vec[a] * M[a + c * size]; - for (a = 0; a < N; a++) { - if (dist > bestd[a]) { - for (d = N - 1; d > a; d--) { - bestd[d] = bestd[d - 1]; - strcpy(bestw[d], bestw[d - 1]); - } - bestd[a] = dist; - strcpy(bestw[a], &vocab[c * max_w]); - break; - } - } - } - for (a = 0; a < N; a++) printf("%50s\t\t%f\n", bestw[a], bestd[a]); - } - return 0; -} diff --git a/contrib/word2vec/word2phrase.c b/contrib/word2vec/word2phrase.c deleted file mode 100644 index 24238bc5b..000000000 --- a/contrib/word2vec/word2phrase.c +++ /dev/null @@ -1,292 +0,0 @@ -// Copyright 2013 Google Inc. All Rights Reserved. -// -// Licensed under the Apache License, Version 2.0 (the "License"); -// you may not use this file except in compliance with the License. -// You may obtain a copy of the License at -// -// http://www.apache.org/licenses/LICENSE-2.0 -// -// Unless required by applicable law or agreed to in writing, software -// distributed under the License is distributed on an "AS IS" BASIS, -// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -// See the License for the specific language governing permissions and -// limitations under the License. - -#include <stdio.h> -#include <stdlib.h> -#include <string.h> -#include <math.h> -#include <pthread.h> - -#define MAX_STRING 60 - -const int vocab_hash_size = 500000000; // Maximum 500M entries in the vocabulary - -typedef float real; // Precision of float numbers - -struct vocab_word { - long long cn; - char *word; -}; - -char train_file[MAX_STRING], output_file[MAX_STRING]; -struct vocab_word *vocab; -int debug_mode = 2, min_count = 5, *vocab_hash, min_reduce = 1; -long long vocab_max_size = 10000, vocab_size = 0; -long long train_words = 0; -real threshold = 100; - -unsigned long long next_random = 1; - -// Reads a single word from a file, assuming space + tab + EOL to be word boundaries -void ReadWord(char *word, FILE *fin) { - int a = 0, ch; - while (!feof(fin)) { - ch = fgetc(fin); - if (ch == 13) continue; - if ((ch == ' ') || (ch == '\t') || (ch == '\n')) { - if (a > 0) { - if (ch == '\n') ungetc(ch, fin); - break; - } - if (ch == '\n') { - strcpy(word, (char *)"</s>"); - return; - } else continue; - } - word[a] = ch; - a++; - if (a >= MAX_STRING - 1) a--; // Truncate too long words - } - word[a] = 0; -} - -// Returns hash value of a word -int GetWordHash(char *word) { - unsigned long long a, hash = 1; - for (a = 0; a < strlen(word); a++) hash = hash * 257 + word[a]; - hash = hash % vocab_hash_size; - return hash; -} - -// Returns position of a word in the vocabulary; if the word is not found, returns -1 -int SearchVocab(char *word) { - unsigned int hash = GetWordHash(word); - while (1) { - if (vocab_hash[hash] == -1) return -1; - if (!strcmp(word, vocab[vocab_hash[hash]].word)) return vocab_hash[hash]; - hash = (hash + 1) % vocab_hash_size; - } - return -1; -} - -// Reads a word and returns its index in the vocabulary -int ReadWordIndex(FILE *fin) { - char word[MAX_STRING]; - ReadWord(word, fin); - if (feof(fin)) return -1; - return SearchVocab(word); -} - -// Adds a word to the vocabulary -int AddWordToVocab(char *word) { - unsigned int hash, length = strlen(word) + 1; - if (length > MAX_STRING) length = MAX_STRING; - vocab[vocab_size].word = (char *)calloc(length, sizeof(char)); - strcpy(vocab[vocab_size].word, word); - vocab[vocab_size].cn = 0; - vocab_size++; - // Reallocate memory if needed - if (vocab_size + 2 >= vocab_max_size) { - vocab_max_size += 10000; - vocab=(struct vocab_word *)realloc(vocab, vocab_max_size * sizeof(struct vocab_word)); - } - hash = GetWordHash(word); - while (vocab_hash[hash] != -1) hash = (hash + 1) % vocab_hash_size; - vocab_hash[hash]=vocab_size - 1; - return vocab_size - 1; -} - -// Used later for sorting by word counts -int VocabCompare(const void *a, const void *b) { - return ((struct vocab_word *)b)->cn - ((struct vocab_word *)a)->cn; -} - -// Sorts the vocabulary by frequency using word counts -void SortVocab() { - int a; - unsigned int hash; - // Sort the vocabulary and keep </s> at the first position - qsort(&vocab[1], vocab_size - 1, sizeof(struct vocab_word), VocabCompare); - for (a = 0; a < vocab_hash_size; a++) vocab_hash[a] = -1; - for (a = 0; a < vocab_size; a++) { - // Words occuring less than min_count times will be discarded from the vocab - if (vocab[a].cn < min_count) { - vocab_size--; - free(vocab[vocab_size].word); - } else { - // Hash will be re-computed, as after the sorting it is not actual - hash = GetWordHash(vocab[a].word); - while (vocab_hash[hash] != -1) hash = (hash + 1) % vocab_hash_size; - vocab_hash[hash] = a; - } - } - vocab = (struct vocab_word *)realloc(vocab, vocab_size * sizeof(struct vocab_word)); -} - -// Reduces the vocabulary by removing infrequent tokens -void ReduceVocab() { - int a, b = 0; - unsigned int hash; - for (a = 0; a < vocab_size; a++) if (vocab[a].cn > min_reduce) { - vocab[b].cn = vocab[a].cn; - vocab[b].word = vocab[a].word; - b++; - } else free(vocab[a].word); - vocab_size = b; - for (a = 0; a < vocab_hash_size; a++) vocab_hash[a] = -1; - for (a = 0; a < vocab_size; a++) { - // Hash will be re-computed, as it is not actual - hash = GetWordHash(vocab[a].word); - while (vocab_hash[hash] != -1) hash = (hash + 1) % vocab_hash_size; - vocab_hash[hash] = a; - } - fflush(stdout); - min_reduce++; -} - -void LearnVocabFromTrainFile() { - char word[MAX_STRING], last_word[MAX_STRING], bigram_word[MAX_STRING * 2]; - FILE *fin; - long long a, i, start = 1; - for (a = 0; a < vocab_hash_size; a++) vocab_hash[a] = -1; - fin = fopen(train_file, "rb"); - if (fin == NULL) { - printf("ERROR: training data file not found!\n"); - exit(1); - } - vocab_size = 0; - AddWordToVocab((char *)"</s>"); - while (1) { - ReadWord(word, fin); - if (feof(fin)) break; - if (!strcmp(word, "</s>")) { - start = 1; - continue; - } else start = 0; - train_words++; - if ((debug_mode > 1) && (train_words % 100000 == 0)) { - printf("Words processed: %lldK Vocab size: %lldK %c", train_words / 1000, vocab_size / 1000, 13); - fflush(stdout); - } - i = SearchVocab(word); - if (i == -1) { - a = AddWordToVocab(word); - vocab[a].cn = 1; - } else vocab[i].cn++; - if (start) continue; - sprintf(bigram_word, "%s_%s", last_word, word); - bigram_word[MAX_STRING - 1] = 0; - strcpy(last_word, word); - i = SearchVocab(bigram_word); - if (i == -1) { - a = AddWordToVocab(bigram_word); - vocab[a].cn = 1; - } else vocab[i].cn++; - if (vocab_size > vocab_hash_size * 0.7) ReduceVocab(); - } - SortVocab(); - if (debug_mode > 0) { - printf("\nVocab size (unigrams + bigrams): %lld\n", vocab_size); - printf("Words in train file: %lld\n", train_words); - } - fclose(fin); -} - -void TrainModel() { - long long pa = 0, pb = 0, pab = 0, oov, i, li = -1, cn = 0; - char word[MAX_STRING], last_word[MAX_STRING], bigram_word[MAX_STRING * 2]; - real score; - FILE *fo, *fin; - printf("Starting training using file %s\n", train_file); - LearnVocabFromTrainFile(); - fin = fopen(train_file, "rb"); - fo = fopen(output_file, "wb"); - word[0] = 0; - while (1) { - strcpy(last_word, word); - ReadWord(word, fin); - if (feof(fin)) break; - if (!strcmp(word, "</s>")) { - fprintf(fo, "\n"); - continue; - } - cn++; - if ((debug_mode > 1) && (cn % 100000 == 0)) { - printf("Words written: %lldK%c", cn / 1000, 13); - fflush(stdout); - } - oov = 0; - i = SearchVocab(word); - if (i == -1) oov = 1; else pb = vocab[i].cn; - if (li == -1) oov = 1; - li = i; - sprintf(bigram_word, "%s_%s", last_word, word); - bigram_word[MAX_STRING - 1] = 0; - i = SearchVocab(bigram_word); - if (i == -1) oov = 1; else pab = vocab[i].cn; - if (pa < min_count) oov = 1; - if (pb < min_count) oov = 1; - if (oov) score = 0; else score = (pab - min_count) / (real)pa / (real)pb * (real)train_words; - if (score > threshold) { - fprintf(fo, "_%s", word); - pb = 0; - } else fprintf(fo, " %s", word); - pa = pb; - } - fclose(fo); - fclose(fin); -} - -int ArgPos(char *str, int argc, char **argv) { - int a; - for (a = 1; a < argc; a++) if (!strcmp(str, argv[a])) { - if (a == argc - 1) { - printf("Argument missing for %s\n", str); - exit(1); - } - return a; - } - return -1; -} - -int main(int argc, char **argv) { - int i; - if (argc == 1) { - printf("WORD2PHRASE tool v0.1a\n\n"); - printf("Options:\n"); - printf("Parameters for training:\n"); - printf("\t-train <file>\n"); - printf("\t\tUse text data from <file> to train the model\n"); - printf("\t-output <file>\n"); - printf("\t\tUse <file> to save the resulting word vectors / word clusters / phrases\n"); - printf("\t-min-count <int>\n"); - printf("\t\tThis will discard words that appear less than <int> times; default is 5\n"); - printf("\t-threshold <float>\n"); - printf("\t\t The <float> value represents threshold for forming the phrases (higher means less phrases); default 100\n"); - printf("\t-debug <int>\n"); - printf("\t\tSet the debug mode (default = 2 = more info during training)\n"); - printf("\nExamples:\n"); - printf("./word2phrase -train text.txt -output phrases.txt -threshold 100 -debug 2\n\n"); - return 0; - } - if ((i = ArgPos((char *)"-train", argc, argv)) > 0) strcpy(train_file, argv[i + 1]); - if ((i = ArgPos((char *)"-debug", argc, argv)) > 0) debug_mode = atoi(argv[i + 1]); - if ((i = ArgPos((char *)"-output", argc, argv)) > 0) strcpy(output_file, argv[i + 1]); - if ((i = ArgPos((char *)"-min-count", argc, argv)) > 0) min_count = atoi(argv[i + 1]); - if ((i = ArgPos((char *)"-threshold", argc, argv)) > 0) threshold = atof(argv[i + 1]); - vocab = (struct vocab_word *)calloc(vocab_max_size, sizeof(struct vocab_word)); - vocab_hash = (int *)calloc(vocab_hash_size, sizeof(int)); - TrainModel(); - return 0; -} |