Преобразование файла word2vec bin в текст
С сайта word2vec я могу скачать GoogleNews-vectors-negative300.bin.gz. Файл .bin(около 3,4 ГБ) - это двоичный формат, который мне не пригодится. Томас Миколов уверяет нас в том, что "должно быть довольно просто преобразовать двоичный формат в текстовый формат (хотя это займет больше места на диске). Проверьте код на дистанционном инструменте, это довольно тривиально для чтения двоичного файла". К сожалению, я не знаю достаточно C, чтобы понять http://word2vec.googlecode.com/svn/trunk/distance.c.
Предположительно gensim может это сделать, но все обучающие материалы, которые я нашел, похоже, касаются преобразования текста, а не другого.
Может ли кто-нибудь предложить изменения кода C или инструкции по генерации текста gensim?
Ответы
Ответ 1
В списке рассылки word2vec-toolkit Thomas Mensink предоставил ответ в виде небольшой программы на C, которая преобразует файл .bin печатать. Это модификация файла distance.c. Я заменил исходное расстояние .c на код Thomas ниже и перестроил word2vec (make clean; make) и переименовал скомпилированное расстояние в readbin. Затем ./readbin vector.bin
создаст текстовую версию vector.bin.
// 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 file_name[max_size];
float len;
long long words, size, a, b;
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));
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);
//Code added by Thomas Mensink
//output the vectors of the binary format in text
printf("%lld %lld #File: %s\n",words,size,file_name);
for (a = 0; a < words; a++){
printf("%s ",&vocab[a * max_w]);
for (b = 0; b< size; b++){ printf("%f ",M[a*size + b]); }
printf("\b\b\n");
}
return 0;
}
Я удалил "\ b\b" из printf
.
Кстати, полученный текстовый файл все еще содержал текстовое слово и некоторые ненужные пробелы, которые я не хотел для некоторых численных вычислений. Я удалил начальный текстовый столбец и конечный пустую строку из каждой строки с помощью команд bash.
cut --complement -d ' ' -f 1 GoogleNews-vectors-negative300.txt > GoogleNews-vectors-negative300_tuples-only.txt
sed 's/ $//' GoogleNews-vectors-negative300_tuples-only.txt
Ответ 2
Я использую этот код для загрузки двоичной модели, затем сохраняю модель в текстовый файл,
from gensim.models.keyedvectors import KeyedVectors
model = KeyedVectors.load_word2vec_format('path/to/GoogleNews-vectors-negative300.bin', binary=True)
model.save_word2vec_format('path/to/GoogleNews-vectors-negative300.txt', binary=False)
Ссылки: API и nullege.
Примечание:
Выше код предназначен для новой версии gensim. Для предыдущей версии я использовал этот код:
from gensim.models import word2vec
model = word2vec.Word2Vec.load_word2vec_format('path/to/GoogleNews-vectors-negative300.bin', binary=True)
model.save_word2vec_format('path/to/GoogleNews-vectors-negative300.txt', binary=False)
Ответ 3
формат - это бинарный формат с плавающей запятой с двойной точностью IEEE 754: binary32
http://en.wikipedia.org/wiki/Single-precision_floating-point_format
Они используют little-endian.
Сделайте пример:
- Первая строка представляет собой строковый формат: "3000000 300\n" (vocabSize &
vecSize, getByte до байта == '\n')
-
Следующая строка включает словарный запас
сначала, а затем (300 * 4 байта значения поплавка, 4 байта для каждого
измерение):
getByte till byte==32 (space). (60 47 115 62 32 => <\s>[space])
-
то каждый следующий 4 байта будет представлять одно число с плавающей запятой
следующий 4 байт: 0 0 -108 58 = > 0,001129150390625.
Вы можете проверить ссылку на wikipedia, чтобы посмотреть, как, позвольте мне сделать это в качестве примера:
(little-endian → обратный порядок) 00111010 10010100 00000000 00000000
- first - знак bit = > sign = 1 (else = -1)
- next 8 bits = > 117 = > exp = 2 ^ (117-127)
- next 23 bits = > pre = 0 * 2 ^ (- 1) + 0 * 2 ^ (- 2) + 1 * 2 ^ (- 3) + 1 * 2 ^ (- 5)
value = sign * exp * pre
Ответ 4
Вы можете загрузить двоичный файл в word2vec, а затем сохранить текстовую версию следующим образом:
from gensim.models import word2vec
model = word2vec.Word2Vec.load_word2vec_format('Path/to/GoogleNews-vectors-negative300.bin', binary=True)
model.save("file.txt")
`
Ответ 5
Я использую gensim для работы с GoogleNews-vector-negative300.bin, и при загрузке модели я включаю флаг binary = True
.
from gensim import word2vec
model = word2vec.Word2Vec.load_word2vec_format('Path/to/GoogleNews-vectors-negative300.bin', binary=True)
Кажется, работает нормально.
Ответ 6
У меня была аналогичная проблема, я хотел получить выходные файлы bin/non-bin (gensim) как CSV.
вот код, который делает это на python, он предполагает, что у вас установлен gensim:
https://gist.github.com/dav009/10a742de43246210f3ba
Ответ 7
convertvec - небольшой инструмент для преобразования векторов между различными форматами для библиотеки word2vec.
Преобразование векторов из двоичного в обычный текст:
./convertvec bin2txt input.bin output.txt
Преобразование векторов из обычного текста в двоичный:
./convertvec txt2bin input.txt output.bin
Ответ 8
Вот код, который я использую:
import codecs
from gensim.models import Word2Vec
def main():
path_to_model = 'GoogleNews-vectors-negative300.bin'
output_file = 'GoogleNews-vectors-negative300_test.txt'
export_to_file(path_to_model, output_file)
def export_to_file(path_to_model, output_file):
output = codecs.open(output_file, 'w' , 'utf-8')
model = Word2Vec.load_word2vec_format(path_to_model, binary=True)
print('done loading Word2Vec')
vocab = model.vocab
for mid in vocab:
#print(model[mid])
#print(mid)
vector = list()
for dimension in model[mid]:
vector.append(str(dimension))
#line = { "mid": mid, "vector": vector }
vector_str = ",".join(vector)
line = mid + "\t" + vector_str
#line = json.dumps(line)
output.write(line + "\n")
output.close()
if __name__ == "__main__":
main()
#cProfile.run('main()') # if you want to do some profiling
Ответ 9
Если вы получаете сообщение об ошибке:
ImportError: No module named models.word2vec
то это потому, что было обновление API. Это будет работать:
from gensim.models.keyedvectors import KeyedVectors
model = KeyedVectors.load_word2vec_format('./GoogleNews-vectors-negative300.bin', binary=True)
model.save_word2vec_format('./GoogleNews-vectors-negative300.txt', binary=False)
Ответ 10
Просто быстрое обновление, так как теперь есть более простой способ.
Если вы используете word2vec
из https://github.com/dav/word2vec, существует дополнительная опция -binary
, которая принимает 1
для генерации двоичного файла или 0
для создания текстового файла. Этот пример исходит от demo-word.sh
в репо:
time ./word2vec -train text8 -output vectors.bin -cbow 1 -size 200 -window 8 -negative 25 -hs 0 -sample 1e-4 -threads 20 -binary 0 -iter 15