#!/usr/bin/env python # -*- coding: utf-8 -*- # # Author: Shiva Manne # Copyright (C) 2018 RaRe Technologies s.r.o. # Licensed under the GNU LGPL v2.1 - http://www.gnu.org/licenses/lgpl.html """This module implements the word2vec family of algorithms, using highly optimized C routines, data streaming and Pythonic interfaces. The word2vec algorithms include skip-gram and CBOW models, using either hierarchical softmax or negative sampling: `Tomas Mikolov et al: Efficient Estimation of Word Representations in Vector Space `_, `Tomas Mikolov et al: Distributed Representations of Words and Phrases and their Compositionality `_. Other embeddings ================ There are more ways to train word vectors in Gensim than just Word2Vec. See also :class:`~gensim.models.doc2vec.Doc2Vec`, :class:`~gensim.models.fasttext.FastText` and wrappers for :class:`~gensim.models.wrappers.VarEmbed` and :class:`~gensim.models.wrappers.WordRank`. The training algorithms were originally ported from the C package https://code.google.com/p/word2vec/ and extended with additional functionality and optimizations over the years. For a tutorial on Gensim word2vec, with an interactive web app trained on GoogleNews, visit https://rare-technologies.com/word2vec-tutorial/. **Make sure you have a C compiler before installing Gensim, to use the optimized word2vec routines** (70x speedup compared to plain NumPy implementation, https://rare-technologies.com/parallelizing-word2vec-in-python/. Usage examples ============== Initialize a model with e.g. >>> from gensim.test.utils import common_texts, get_tmpfile >>> from gensim.models import Word2Vec >>> >>> path = get_tmpfile("word2vec.model") >>> >>> model = Word2Vec(common_texts, size=100, window=5, min_count=1, workers=4) >>> model.save("word2vec.model") The training is streamed, meaning `sentences` can be a generator, reading input data from disk on-the-fly, without loading the entire corpus into RAM. It also means you can continue training the model later >>> model = Word2Vec.load("word2vec.model") >>> model.train([["hello", "world"]], total_examples=1, epochs=1) (0, 2) The trained word vectors are stored in a :class:`~gensim.models.KeyedVectors` instance in `model.wv`: >>> vector = model.wv['computer'] # numpy vector of a word The reason for separating the trained vectors into `KeyedVectors` is that if you don't need the full model state any more (don't need to continue training), the state can discarded, resulting in a much smaller and faster object that can be mmapped for lightning fast loading and sharing the vectors in RAM between processes:: >>> from gensim.models import KeyedVectors >>> >>> path = get_tmpfile("wordvectors.kv") >>> >>> model.wv.save(path) >>> wv = KeyedVectors.load("model.wv", mmap='r') >>> vector = wv['computer'] # numpy vector of a word Gensim can also load word vectors in the "word2vec C format", as this :class:`~gensim.models.KeyedVectors` instance:: >>> from gensim.test.utils import datapath >>> >>> wv_from_text = KeyedVectors.load_word2vec_format(datapath('word2vec_pre_kv_c'), binary=False) # C text format >>> wv_from_bin = KeyedVectors.load_word2vec_format(datapath("euclidean_vectors.bin"), binary=True) # C binary format It is impossible to continue training the vectors loaded from the C format because the hidden weights, vocabulary frequencies and the binary tree are missing. To continue training, you'll need the full :class:`~gensim.models.word2vec.Word2Vec` object state, as stored by :meth:`~gensim.models.word2vec.Word2Vec.save`, not just the :class:`~gensim.models.keyedvectors.KeyedVectors`. You can perform various NLP word tasks with a trained model. Some of them are already built-in - you can see it in :mod:`gensim.models.keyedvectors`. If you're finished training a model (i.e. no more updates, only querying), you can switch to the :class:`~gensim.models.KeyedVectors` instance >>> word_vectors = model.wv >>> del model to trim unneeded model state = use much less RAM and allow fast loading and memory sharing (mmap). Note that there is a :mod:`gensim.models.phrases` module which lets you automatically detect phrases longer than one word. Using phrases, you can learn a word2vec model where "words" are actually multiword expressions, such as `new_york_times` or `financial_crisis`: >>> from gensim.test.utils import common_texts >>> from gensim.models import Phrases >>> >>> bigram_transformer = Phrases(common_texts) >>> model = Word2Vec(bigram_transformer[common_texts], min_count=1) """ from __future__ import division # py3 "true division" import logging import sys import os import heapq from timeit import default_timer from copy import deepcopy from collections import defaultdict import threading import itertools import warnings from gensim.utils import keep_vocab_item, call_on_class_only from gensim.models.keyedvectors import Vocab, Word2VecKeyedVectors from gensim.models.base_any2vec import BaseWordEmbeddingsModel try: from queue import Queue, Empty except ImportError: from Queue import Queue, Empty from numpy import exp, dot, zeros, random, dtype, float32 as REAL,\ uint32, seterr, array, uint8, vstack, fromstring, sqrt,\ empty, sum as np_sum, ones, logaddexp, log, outer from scipy.special import expit from gensim import utils, matutils # utility fnc for pickling, common scipy operations etc from gensim.utils import deprecated from six import iteritems, itervalues, string_types from six.moves import xrange logger = logging.getLogger(__name__) try: from gensim.models.word2vec_inner import train_batch_sg, train_batch_cbow from gensim.models.word2vec_inner import score_sentence_sg, score_sentence_cbow from gensim.models.word2vec_inner import FAST_VERSION, MAX_WORDS_IN_BATCH except ImportError: # failed... fall back to plain numpy (20-80x slower training than the above) FAST_VERSION = -1 MAX_WORDS_IN_BATCH = 10000 def train_batch_sg(model, sentences, alpha, work=None, compute_loss=False): """Update skip-gram model by training on a sequence of sentences. Called internally from :meth:`~gensim.models.word2vec.Word2Vec.train`. Warnings -------- This is the non-optimized, pure Python version. If you have a C compiler, Gensim will use an optimized code path from :mod:`gensim.models.word2vec_inner` instead. Parameters ---------- model : :class:`~gensim.models.word2Vec.Word2Vec` The Word2Vec model instance to train. sentences : iterable of list of str The corpus used to train the model. alpha : float The learning rate work : object, optional Unused. compute_loss : bool, optional Whether or not the training loss should be computed in this batch. Returns ------- int Number of words in the vocabulary actually used for training (that already existed in the vocabulary and were not discarded by negative sampling). """ result = 0 for sentence in sentences: word_vocabs = [model.wv.vocab[w] for w in sentence if w in model.wv.vocab and model.wv.vocab[w].sample_int > model.random.rand() * 2 ** 32] for pos, word in enumerate(word_vocabs): reduced_window = model.random.randint(model.window) # `b` in the original word2vec code # now go over all words from the (reduced) window, predicting each one in turn start = max(0, pos - model.window + reduced_window) for pos2, word2 in enumerate(word_vocabs[start:(pos + model.window + 1 - reduced_window)], start): # don't train on the `word` itself if pos2 != pos: train_sg_pair( model, model.wv.index2word[word.index], word2.index, alpha, compute_loss=compute_loss ) result += len(word_vocabs) return result def train_batch_cbow(model, sentences, alpha, work=None, neu1=None, compute_loss=False): """Update CBOW model by training on a sequence of sentences. Called internally from :meth:`~gensim.models.word2vec.Word2Vec.train`. Warnings -------- This is the non-optimized, pure Python version. If you have a C compiler, Gensim will use an optimized code path from :mod:`gensim.models.word2vec_inner` instead. Parameters ---------- model : :class:`~gensim.models.word2vec.Word2Vec` The Word2Vec model instance to train. sentences : iterable of list of str The corpus used to train the model. alpha : float The learning rate work : object, optional Unused. neu1 : object, optional Unused. compute_loss : bool, optional Whether or not the training loss should be computed in this batch. Returns ------- int Number of words in the vocabulary actually used for training (that already existed in the vocabulary and were not discarded by negative sampling). """ result = 0 for sentence in sentences: word_vocabs = [ model.wv.vocab[w] for w in sentence if w in model.wv.vocab and model.wv.vocab[w].sample_int > model.random.rand() * 2 ** 32 ] for pos, word in enumerate(word_vocabs): reduced_window = model.random.randint(model.window) # `b` in the original word2vec code start = max(0, pos - model.window + reduced_window) window_pos = enumerate(word_vocabs[start:(pos + model.window + 1 - reduced_window)], start) word2_indices = [word2.index for pos2, word2 in window_pos if (word2 is not None and pos2 != pos)] l1 = np_sum(model.wv.syn0[word2_indices], axis=0) # 1 x vector_size if word2_indices and model.cbow_mean: l1 /= len(word2_indices) train_cbow_pair(model, word, word2_indices, l1, alpha, compute_loss=compute_loss) result += len(word_vocabs) return result def score_sentence_sg(model, sentence, work=None): """Obtain likelihood score for a single sentence in a fitted skip-gram representation. Notes ----- This is the non-optimized, pure Python version. If you have a C compiler, Gensim will use an optimized code path from :mod:`gensim.models.word2vec_inner` instead. Parameters ---------- model : :class:`~gensim.models.word2vec.Word2Vec` The trained model. It **MUST** have been trained using hierarchical softmax and the skip-gram algorithm. sentence : list of str The words comprising the sentence to be scored. work : object, optional Unused. For interface compatibility only. Returns ------- float The probability assigned to this sentence by the Skip-Gram model. """ log_prob_sentence = 0.0 if model.negative: raise RuntimeError("scoring is only available for HS=True") word_vocabs = [model.wv.vocab[w] for w in sentence if w in model.wv.vocab] for pos, word in enumerate(word_vocabs): if word is None: continue # OOV word in the input sentence => skip # now go over all words from the window, predicting each one in turn start = max(0, pos - model.window) for pos2, word2 in enumerate(word_vocabs[start: pos + model.window + 1], start): # don't train on OOV words and on the `word` itself if word2 is not None and pos2 != pos: log_prob_sentence += score_sg_pair(model, word, word2) return log_prob_sentence def score_sentence_cbow(model, sentence, work=None, neu1=None): """Obtain likelihood score for a single sentence in a fitted CBOW representation. Notes ----- This is the non-optimized, pure Python version. If you have a C compiler, Gensim will use an optimized code path from :mod:`gensim.models.word2vec_inner` instead. Parameters ---------- model : :class:`~gensim.models.word2vec.Word2Vec` The trained model. It **MUST** have been trained using hierarchical softmax and the CBOW algorithm. sentence : list of str The words comprising the sentence to be scored. work : object, optional Unused. For interface compatibility only. neu1 : object, optional Unused. For interface compatibility only. Returns ------- float The probability assigned to this sentence by the CBOW model. """ log_prob_sentence = 0.0 if model.negative: raise RuntimeError("scoring is only available for HS=True") word_vocabs = [model.wv.vocab[w] for w in sentence if w in model.wv.vocab] for pos, word in enumerate(word_vocabs): if word is None: continue # OOV word in the input sentence => skip start = max(0, pos - model.window) window_pos = enumerate(word_vocabs[start:(pos + model.window + 1)], start) word2_indices = [word2.index for pos2, word2 in window_pos if (word2 is not None and pos2 != pos)] l1 = np_sum(model.wv.syn0[word2_indices], axis=0) # 1 x layer1_size if word2_indices and model.cbow_mean: l1 /= len(word2_indices) log_prob_sentence += score_cbow_pair(model, word, l1) return log_prob_sentence def train_sg_pair(model, word, context_index, alpha, learn_vectors=True, learn_hidden=True, context_vectors=None, context_locks=None, compute_loss=False, is_ft=False): """Train the passed model instance on a word and its context, using the Skip-gram algorithm. Parameters ---------- model : :class:`~gensim.models.word2vec.Word2Vec` The model to be trained. word : str The label (predicted) word. context_index : list of int The vocabulary indices of the words in the context. alpha : float Learning rate. learn_vectors : bool, optional Whether the vectors should be updated. learn_hidden : bool, optional Whether the weights of the hidden layer should be updated. context_vectors : list of list of float, optional Vector representations of the words in the context. If None, these will be retrieved from the model. context_locks : list of float, optional The lock factors for each word in the context. compute_loss : bool, optional Whether or not the training loss should be computed. is_ft : bool, optional If True, weights will be computed using `model.wv.syn0_vocab` and `model.wv.syn0_ngrams` instead of `model.wv.syn0`. Returns ------- numpy.ndarray Error vector to be back-propagated. """ if context_vectors is None: if is_ft: context_vectors_vocab = model.wv.syn0_vocab context_vectors_ngrams = model.wv.syn0_ngrams else: context_vectors = model.wv.syn0 if context_locks is None: if is_ft: context_locks_vocab = model.syn0_vocab_lockf context_locks_ngrams = model.syn0_ngrams_lockf else: context_locks = model.syn0_lockf if word not in model.wv.vocab: return predict_word = model.wv.vocab[word] # target word (NN output) if is_ft: l1_vocab = context_vectors_vocab[context_index[0]] l1_ngrams = np_sum(context_vectors_ngrams[context_index[1:]], axis=0) if context_index: l1 = np_sum([l1_vocab, l1_ngrams], axis=0) / len(context_index) else: l1 = context_vectors[context_index] # input word (NN input/projection layer) lock_factor = context_locks[context_index] neu1e = zeros(l1.shape) if model.hs: # work on the entire tree at once, to push as much work into numpy's C routines as possible (performance) l2a = deepcopy(model.syn1[predict_word.point]) # 2d matrix, codelen x layer1_size prod_term = dot(l1, l2a.T) fa = expit(prod_term) # propagate hidden -> output ga = (1 - predict_word.code - fa) * alpha # vector of error gradients multiplied by the learning rate if learn_hidden: model.syn1[predict_word.point] += outer(ga, l1) # learn hidden -> output neu1e += dot(ga, l2a) # save error # loss component corresponding to hierarchical softmax if compute_loss: sgn = (-1.0) ** predict_word.code # `ch` function, 0 -> 1, 1 -> -1 lprob = -log(expit(-sgn * prod_term)) model.running_training_loss += sum(lprob) if model.negative: # use this word (label = 1) + `negative` other random words not from this sentence (label = 0) word_indices = [predict_word.index] while len(word_indices) < model.negative + 1: w = model.cum_table.searchsorted(model.random.randint(model.cum_table[-1])) if w != predict_word.index: word_indices.append(w) l2b = model.syn1neg[word_indices] # 2d matrix, k+1 x layer1_size prod_term = dot(l1, l2b.T) fb = expit(prod_term) # propagate hidden -> output gb = (model.neg_labels - fb) * alpha # vector of error gradients multiplied by the learning rate if learn_hidden: model.syn1neg[word_indices] += outer(gb, l1) # learn hidden -> output neu1e += dot(gb, l2b) # save error # loss component corresponding to negative sampling if compute_loss: model.running_training_loss -= sum(log(expit(-1 * prod_term[1:]))) # for the sampled words model.running_training_loss -= log(expit(prod_term[0])) # for the output word if learn_vectors: if is_ft: model.wv.syn0_vocab[context_index[0]] += neu1e * context_locks_vocab[context_index[0]] for i in context_index[1:]: model.wv.syn0_ngrams[i] += neu1e * context_locks_ngrams[i] else: l1 += neu1e * lock_factor # learn input -> hidden (mutates model.wv.syn0[word2.index], if that is l1) return neu1e def train_cbow_pair(model, word, input_word_indices, l1, alpha, learn_vectors=True, learn_hidden=True, compute_loss=False, context_vectors=None, context_locks=None, is_ft=False): """Train the passed model instance on a word and its context, using the CBOW algorithm. Parameters ---------- model : :class:`~gensim.models.word2vec.Word2Vec` The model to be trained. word : str The label (predicted) word. input_word_indices : list of int The vocabulary indices of the words in the context. l1 : list of float Vector representation of the label word. alpha : float Learning rate. learn_vectors : bool, optional Whether the vectors should be updated. learn_hidden : bool, optional Whether the weights of the hidden layer should be updated. compute_loss : bool, optional Whether or not the training loss should be computed. context_vectors : list of list of float, optional Vector representations of the words in the context. If None, these will be retrieved from the model. context_locks : list of float, optional The lock factors for each word in the context. is_ft : bool, optional If True, weights will be computed using `model.wv.syn0_vocab` and `model.wv.syn0_ngrams` instead of `model.wv.syn0`. Returns ------- numpy.ndarray Error vector to be back-propagated. """ if context_vectors is None: if is_ft: context_vectors_vocab = model.wv.syn0_vocab context_vectors_ngrams = model.wv.syn0_ngrams else: context_vectors = model.wv.syn0 if context_locks is None: if is_ft: context_locks_vocab = model.syn0_vocab_lockf context_locks_ngrams = model.syn0_ngrams_lockf else: context_locks = model.syn0_lockf neu1e = zeros(l1.shape) if model.hs: l2a = model.syn1[word.point] # 2d matrix, codelen x layer1_size prod_term = dot(l1, l2a.T) fa = expit(prod_term) # propagate hidden -> output ga = (1. - word.code - fa) * alpha # vector of error gradients multiplied by the learning rate if learn_hidden: model.syn1[word.point] += outer(ga, l1) # learn hidden -> output neu1e += dot(ga, l2a) # save error # loss component corresponding to hierarchical softmax if compute_loss: sgn = (-1.0) ** word.code # ch function, 0-> 1, 1 -> -1 model.running_training_loss += sum(-log(expit(-sgn * prod_term))) if model.negative: # use this word (label = 1) + `negative` other random words not from this sentence (label = 0) word_indices = [word.index] while len(word_indices) < model.negative + 1: w = model.cum_table.searchsorted(model.random.randint(model.cum_table[-1])) if w != word.index: word_indices.append(w) l2b = model.syn1neg[word_indices] # 2d matrix, k+1 x layer1_size prod_term = dot(l1, l2b.T) fb = expit(prod_term) # propagate hidden -> output gb = (model.neg_labels - fb) * alpha # vector of error gradients multiplied by the learning rate if learn_hidden: model.syn1neg[word_indices] += outer(gb, l1) # learn hidden -> output neu1e += dot(gb, l2b) # save error # loss component corresponding to negative sampling if compute_loss: model.running_training_loss -= sum(log(expit(-1 * prod_term[1:]))) # for the sampled words model.running_training_loss -= log(expit(prod_term[0])) # for the output word if learn_vectors: # learn input -> hidden, here for all words in the window separately if is_ft: if not model.cbow_mean and input_word_indices: neu1e /= (len(input_word_indices[0]) + len(input_word_indices[1])) for i in input_word_indices[0]: context_vectors_vocab[i] += neu1e * context_locks_vocab[i] for i in input_word_indices[1]: context_vectors_ngrams[i] += neu1e * context_locks_ngrams[i] else: if not model.cbow_mean and input_word_indices: neu1e /= len(input_word_indices) for i in input_word_indices: context_vectors[i] += neu1e * context_locks[i] return neu1e def score_sg_pair(model, word, word2): """Score the trained Skip-gram model on a pair of words. Parameters ---------- model : :class:`~gensim.models.word2vec.Word2Vec` The trained model. word : :class:`~gensim.models.keyedvectors.Vocab` Vocabulary representation of the first word. word2 : :class:`~gensim.models.keyedvectors.Vocab` Vocabulary representation of the second word. Returns ------- float Logarithm of the sum of exponentiations of input words. """ l1 = model.wv.syn0[word2.index] l2a = deepcopy(model.syn1[word.point]) # 2d matrix, codelen x layer1_size sgn = (-1.0) ** word.code # ch function, 0-> 1, 1 -> -1 lprob = -logaddexp(0, -sgn * dot(l1, l2a.T)) return sum(lprob) def score_cbow_pair(model, word, l1): """Score the trained CBOW model on a pair of words. Parameters ---------- model : :class:`~gensim.models.word2vec.Word2Vec` The trained model. word : :class:`~gensim.models.keyedvectors.Vocab` Vocabulary representation of the first word. l1 : list of float Vector representation of the second word. Returns ------- float Logarithm of the sum of exponentiations of input words. """ l2a = model.syn1[word.point] # 2d matrix, codelen x layer1_size sgn = (-1.0) ** word.code # ch function, 0-> 1, 1 -> -1 lprob = -logaddexp(0, -sgn * dot(l1, l2a.T)) return sum(lprob) class Word2Vec(BaseWordEmbeddingsModel): """Train, use and evaluate neural networks described in https://code.google.com/p/word2vec/. Once you're finished training a model (=no more updates, only querying) store and use only the :class:`~gensim.models.keyedvectors.KeyedVectors` instance in `self.wv` to reduce memory. The model can be stored/loaded via its :meth:`~gensim.models.word2vec.Word2Vec.save` and :meth:`~gensim.models.word2vec.Word2Vec.load` methods. The trained word vectors can also be stored/loaded from a format compatible with the original word2vec implementation via `self.wv.save_word2vec_format` and :meth:`gensim.models.keyedvectors.KeyedVectors.load_word2vec_format`. Some important attributes are the following: Attributes ---------- wv : :class:`~gensim.models.keyedvectors.Word2VecKeyedVectors` This object essentially contains the mapping between words and embeddings. After training, it can be used directly to query those embeddings in various ways. See the module level docstring for examples. vocabulary : :class:'~gensim.models.word2vec.Word2VecVocab' This object represents the vocabulary (sometimes called Dictionary in gensim) of the model. Besides keeping track of all unique words, this object provides extra functionality, such as constructing a huffman tree (frequent words are closer to the root), or discarding extremely rare words. trainables : :class:`~gensim.models.word2vec.Word2VecTrainables` This object represents the inner shallow neural network used to train the embeddings. The semantics of the network differ slightly in the two available training modes (CBOW or SG) but you can think of it as a NN with a single projection and hidden layer which we train on the corpus. The weights are then used as our embeddings (which means that the size of the hidden layer is equal to the number of features `self.size`). """ def __init__(self, sentences=None, size=100, alpha=0.025, window=5, min_count=5, max_vocab_size=None, sample=1e-3, seed=1, workers=3, min_alpha=0.0001, sg=0, hs=0, negative=5, ns_exponent=0.75, cbow_mean=1, hashfxn=hash, iter=5, null_word=0, trim_rule=None, sorted_vocab=1, batch_words=MAX_WORDS_IN_BATCH, compute_loss=False, callbacks=(), max_final_vocab=None): """ Parameters ---------- sentences : iterable of iterables, optional The `sentences` iterable can be simply a list of lists of tokens, but for larger corpora, consider an iterable that streams the sentences directly from disk/network. See :class:`~gensim.models.word2vec.BrownCorpus`, :class:`~gensim.models.word2vec.Text8Corpus` or :class:`~gensim.models.word2vec.LineSentence` in :mod:`~gensim.models.word2vec` module for such examples. See also the `tutorial on data streaming in Python `_. If you don't supply `sentences`, the model is left uninitialized -- use if you plan to initialize it in some other way. size : int, optional Dimensionality of the word vectors. window : int, optional Maximum distance between the current and predicted word within a sentence. min_count : int, optional Ignores all words with total frequency lower than this. workers : int, optional Use these many worker threads to train the model (=faster training with multicore machines). sg : {0, 1}, optional Training algorithm: 1 for skip-gram; otherwise CBOW. hs : {0, 1}, optional If 1, hierarchical softmax will be used for model training. If 0, and `negative` is non-zero, negative sampling will be used. negative : int, optional If > 0, negative sampling will be used, the int for negative specifies how many "noise words" should be drawn (usually between 5-20). If set to 0, no negative sampling is used. ns_exponent : float, optional The exponent used to shape the negative sampling distribution. A value of 1.0 samples exactly in proportion to the frequencies, 0.0 samples all words equally, while a negative value samples low-frequency words more than high-frequency words. The popular default value of 0.75 was chosen by the original Word2Vec paper. More recently, in https://arxiv.org/abs/1804.04212, Caselles-Dupré, Lesaint, & Royo-Letelier suggest that other values may perform better for recommendation applications. cbow_mean : {0, 1}, optional If 0, use the sum of the context word vectors. If 1, use the mean, only applies when cbow is used. alpha : float, optional The initial learning rate. min_alpha : float, optional Learning rate will linearly drop to `min_alpha` as training progresses. seed : int, optional Seed for the random number generator. Initial vectors for each word are seeded with a hash of the concatenation of word + `str(seed)`. Note that for a fully deterministically-reproducible run, you must also limit the model to a single worker thread (`workers=1`), to eliminate ordering jitter from OS thread scheduling. (In Python 3, reproducibility between interpreter launches also requires use of the `PYTHONHASHSEED` environment variable to control hash randomization). max_vocab_size : int, optional Limits the RAM during vocabulary building; if there are more unique words than this, then prune the infrequent ones. Every 10 million word types need about 1GB of RAM. Set to `None` for no limit. max_final_vocab : int, optional Limits the vocab to a target vocab size by automatically picking a matching min_count. If the specified min_count is more than the calculated min_count, the specified min_count will be used. Set to `None` if not required. sample : float, optional The threshold for configuring which higher-frequency words are randomly downsampled, useful range is (0, 1e-5). hashfxn : function, optional Hash function to use to randomly initialize weights, for increased training reproducibility. iter : int, optional Number of iterations (epochs) over the corpus. trim_rule : function, optional Vocabulary trimming rule, specifies whether certain words should remain in the vocabulary, be trimmed away, or handled using the default (discard if word count < min_count). Can be None (min_count will be used, look to :func:`~gensim.utils.keep_vocab_item`), or a callable that accepts parameters (word, count, min_count) and returns either :attr:`gensim.utils.RULE_DISCARD`, :attr:`gensim.utils.RULE_KEEP` or :attr:`gensim.utils.RULE_DEFAULT`. The rule, if given, is only used to prune vocabulary during build_vocab() and is not stored as part of the model. The input parameters are of the following types: * `word` (str) - the word we are examining * `count` (int) - the word's frequency count in the corpus * `min_count` (int) - the minimum count threshold. sorted_vocab : {0, 1}, optional If 1, sort the vocabulary by descending frequency before assigning word indexes. See :meth:`~gensim.models.word2vec.Word2VecVocab.sort_vocab()`. batch_words : int, optional Target size (in words) for batches of examples passed to worker threads (and thus cython routines).(Larger batches will be passed if individual texts are longer than 10000 words, but the standard cython code truncates to that maximum.) compute_loss: bool, optional If True, computes and stores loss value which can be retrieved using :meth:`~gensim.models.word2vec.Word2Vec.get_latest_training_loss`. callbacks : iterable of :class:`~gensim.models.callbacks.CallbackAny2Vec`, optional Sequence of callbacks to be executed at specific stages during training. Examples -------- Initialize and train a :class:`~gensim.models.word2vec.Word2Vec` model >>> from gensim.models import Word2Vec >>> sentences = [["cat", "say", "meow"], ["dog", "say", "woof"]] >>> model = Word2Vec(sentences, min_count=1) """ self.max_final_vocab = max_final_vocab self.callbacks = callbacks self.load = call_on_class_only self.wv = Word2VecKeyedVectors(size) self.vocabulary = Word2VecVocab( max_vocab_size=max_vocab_size, min_count=min_count, sample=sample, sorted_vocab=bool(sorted_vocab), null_word=null_word, max_final_vocab=max_final_vocab, ns_exponent=ns_exponent) self.trainables = Word2VecTrainables(seed=seed, vector_size=size, hashfxn=hashfxn) super(Word2Vec, self).__init__( sentences=sentences, workers=workers, vector_size=size, epochs=iter, callbacks=callbacks, batch_words=batch_words, trim_rule=trim_rule, sg=sg, alpha=alpha, window=window, seed=seed, hs=hs, negative=negative, cbow_mean=cbow_mean, min_alpha=min_alpha, compute_loss=compute_loss, fast_version=FAST_VERSION) def _do_train_job(self, sentences, alpha, inits): """Train the model on a single batch of sentences. Parameters ---------- sentences : iterable of list of str Corpus chunk to be used in this training batch. alpha : float The learning rate used in this batch. inits : (np.ndarray, np.ndarray) Each worker threads private work memory. Returns ------- (int, int) 2-tuple (effective word count after ignoring unknown words and sentence length trimming, total word count). """ work, neu1 = inits tally = 0 if self.sg: tally += train_batch_sg(self, sentences, alpha, work, self.compute_loss) else: tally += train_batch_cbow(self, sentences, alpha, work, neu1, self.compute_loss) return tally, self._raw_word_count(sentences) def _clear_post_train(self): """Remove all L2-normalized word vectors from the model.""" self.wv.vectors_norm = None def _set_train_params(self, **kwargs): if 'compute_loss' in kwargs: self.compute_loss = kwargs['compute_loss'] self.running_training_loss = 0 def train(self, sentences, total_examples=None, total_words=None, epochs=None, start_alpha=None, end_alpha=None, word_count=0, queue_factor=2, report_delay=1.0, compute_loss=False, callbacks=()): """Update the model's neural weights from a sequence of sentences. Notes ----- To support linear learning-rate decay from (initial) `alpha` to `min_alpha`, and accurate progress-percentage logging, either `total_examples` (count of sentences) or `total_words` (count of raw words in sentences) **MUST** be provided. If `sentences` is the same corpus that was provided to :meth:`~gensim.models.word2vec.Word2Vec.build_vocab` earlier, you can simply use `total_examples=self.corpus_count`. Warnings -------- To avoid common mistakes around the model's ability to do multiple training passes itself, an explicit `epochs` argument **MUST** be provided. In the common and recommended case where :meth:`~gensim.models.word2vec.Word2Vec.train` is only called once, you can set `epochs=self.iter`. Parameters ---------- sentences : iterable of list of str The `sentences` iterable can be simply a list of lists of tokens, but for larger corpora, consider an iterable that streams the sentences directly from disk/network. See :class:`~gensim.models.word2vec.BrownCorpus`, :class:`~gensim.models.word2vec.Text8Corpus` or :class:`~gensim.models.word2vec.LineSentence` in :mod:`~gensim.models.word2vec` module for such examples. See also the `tutorial on data streaming in Python `_. total_examples : int, optional Count of sentences. Used to decay the `alpha` learning rate. total_words : int, optional Count of raw words in sentences. Used to decay the `alpha` learning rate. epochs : int, optional Number of iterations (epochs) over the corpus. start_alpha : float, optional Initial learning rate. If supplied, replaces the starting `alpha` from the constructor, for this one call to`train()`. Use only if making multiple calls to `train()`, when you want to manage the alpha learning-rate yourself (not recommended). end_alpha : float, optional Final learning rate. Drops linearly from `start_alpha`. If supplied, this replaces the final `min_alpha` from the constructor, for this one call to `train()`. Use only if making multiple calls to `train()`, when you want to manage the alpha learning-rate yourself (not recommended). word_count : int, optional Count of words already trained. Set this to 0 for the usual case of training on all words in sentences. queue_factor : int, optional Multiplier for size of queue (number of workers * queue_factor). report_delay : float, optional Seconds to wait before reporting progress. compute_loss: bool, optional If True, computes and stores loss value which can be retrieved using :meth:`~gensim.models.word2vec.Word2Vec.get_latest_training_loss`. callbacks : iterable of :class:`~gensim.models.callbacks.CallbackAny2Vec`, optional Sequence of callbacks to be executed at specific stages during training. Examples -------- >>> from gensim.models import Word2Vec >>> sentences = [["cat", "say", "meow"], ["dog", "say", "woof"]] >>> >>> model = Word2Vec(min_count=1) >>> model.build_vocab(sentences) # prepare the model vocabulary >>> model.train(sentences, total_examples=model.corpus_count, epochs=model.iter) # train word vectors (1, 30) """ return super(Word2Vec, self).train( sentences, total_examples=total_examples, total_words=total_words, epochs=epochs, start_alpha=start_alpha, end_alpha=end_alpha, word_count=word_count, queue_factor=queue_factor, report_delay=report_delay, compute_loss=compute_loss, callbacks=callbacks) def score(self, sentences, total_sentences=int(1e6), chunksize=100, queue_factor=2, report_delay=1): """Score the log probability for a sequence of sentences. This does not change the fitted model in any way (see :meth:`~gensim.models.word2vec.Word2Vec.train` for that). Gensim has currently only implemented score for the hierarchical softmax scheme, so you need to have run word2vec with `hs=1` and `negative=0` for this to work. Note that you should specify `total_sentences`; you'll run into problems if you ask to score more than this number of sentences but it is inefficient to set the value too high. See the `article by Matt Taddy: "Document Classification by Inversion of Distributed Language Representations" `_ and the `gensim demo `_ for examples of how to use such scores in document classification. Parameters ---------- sentences : iterable of list of str The `sentences` iterable can be simply a list of lists of tokens, but for larger corpora, consider an iterable that streams the sentences directly from disk/network. See :class:`~gensim.models.word2vec.BrownCorpus`, :class:`~gensim.models.word2vec.Text8Corpus` or :class:`~gensim.models.word2vec.LineSentence` in :mod:`~gensim.models.word2vec` module for such examples. total_sentences : int, optional Count of sentences. chunksize : int, optional Chunksize of jobs queue_factor : int, optional Multiplier for size of queue (number of workers * queue_factor). report_delay : float, optional Seconds to wait before reporting progress. """ if FAST_VERSION < 0: warnings.warn( "C extension compilation failed, scoring will be slow. " "Install a C compiler and reinstall gensim for fastness." ) logger.info( "scoring sentences with %i workers on %i vocabulary and %i features, " "using sg=%s hs=%s sample=%s and negative=%s", self.workers, len(self.wv.vocab), self.trainables.layer1_size, self.sg, self.hs, self.vocabulary.sample, self.negative ) if not self.wv.vocab: raise RuntimeError("you must first build vocabulary before scoring new data") if not self.hs: raise RuntimeError( "We have currently only implemented score for the hierarchical softmax scheme, " "so you need to have run word2vec with hs=1 and negative=0 for this to work." ) def worker_loop(): """Compute log probability for each sentence, lifting lists of sentences from the jobs queue.""" work = zeros(1, dtype=REAL) # for sg hs, we actually only need one memory loc (running sum) neu1 = matutils.zeros_aligned(self.trainables.layer1_size, dtype=REAL) while True: job = job_queue.get() if job is None: # signal to finish break ns = 0 for sentence_id, sentence in job: if sentence_id >= total_sentences: break if self.sg: score = score_sentence_sg(self, sentence, work) else: score = score_sentence_cbow(self, sentence, work, neu1) sentence_scores[sentence_id] = score ns += 1 progress_queue.put(ns) # report progress start, next_report = default_timer(), 1.0 # buffer ahead only a limited number of jobs.. this is the reason we can't simply use ThreadPool :( job_queue = Queue(maxsize=queue_factor * self.workers) progress_queue = Queue(maxsize=(queue_factor + 1) * self.workers) workers = [threading.Thread(target=worker_loop) for _ in xrange(self.workers)] for thread in workers: thread.daemon = True # make interrupting the process with ctrl+c easier thread.start() sentence_count = 0 sentence_scores = matutils.zeros_aligned(total_sentences, dtype=REAL) push_done = False done_jobs = 0 jobs_source = enumerate(utils.grouper(enumerate(sentences), chunksize)) # fill jobs queue with (id, sentence) job items while True: try: job_no, items = next(jobs_source) if (job_no - 1) * chunksize > total_sentences: logger.warning( "terminating after %i sentences (set higher total_sentences if you want more).", total_sentences ) job_no -= 1 raise StopIteration() logger.debug("putting job #%i in the queue", job_no) job_queue.put(items) except StopIteration: logger.info("reached end of input; waiting to finish %i outstanding jobs", job_no - done_jobs + 1) for _ in xrange(self.workers): job_queue.put(None) # give the workers heads up that they can finish -- no more work! push_done = True try: while done_jobs < (job_no + 1) or not push_done: ns = progress_queue.get(push_done) # only block after all jobs pushed sentence_count += ns done_jobs += 1 elapsed = default_timer() - start if elapsed >= next_report: logger.info( "PROGRESS: at %.2f%% sentences, %.0f sentences/s", 100.0 * sentence_count, sentence_count / elapsed ) next_report = elapsed + report_delay # don't flood log, wait report_delay seconds else: # loop ended by job count; really done break except Empty: pass # already out of loop; continue to next push elapsed = default_timer() - start self.clear_sims() logger.info( "scoring %i sentences took %.1fs, %.0f sentences/s", sentence_count, elapsed, sentence_count / elapsed ) return sentence_scores[:sentence_count] def clear_sims(self): """Remove all L2-normalized word vectors from the model, to free up memory. You can recompute them later again using the :meth:`~gensim.models.word2vec.Word2Vec.init_sims` method. """ self.wv.vectors_norm = None def intersect_word2vec_format(self, fname, lockf=0.0, binary=False, encoding='utf8', unicode_errors='strict'): """Merge in an input-hidden weight matrix loaded from the original C word2vec-tool format, where it intersects with the current vocabulary. No words are added to the existing vocabulary, but intersecting words adopt the file's weights, and non-intersecting words are left alone. Parameters ---------- fname : str The file path to load the vectors from. lockf : float, optional Lock-factor value to be set for any imported word-vectors; the default value of 0.0 prevents further updating of the vector during subsequent training. Use 1.0 to allow further training updates of merged vectors. binary : bool, optional If True, `fname` is in the binary word2vec C format. encoding : str, optional Encoding of `text` for `unicode` function (python2 only). unicode_errors : str, optional Error handling behaviour, used as parameter for `unicode` function (python2 only). """ overlap_count = 0 logger.info("loading projection weights from %s", fname) with utils.smart_open(fname) as fin: header = utils.to_unicode(fin.readline(), encoding=encoding) vocab_size, vector_size = (int(x) for x in header.split()) # throws for invalid file format if not vector_size == self.wv.vector_size: raise ValueError("incompatible vector size %d in file %s" % (vector_size, fname)) # TOCONSIDER: maybe mismatched vectors still useful enough to merge (truncating/padding)? if binary: binary_len = dtype(REAL).itemsize * vector_size for _ in xrange(vocab_size): # mixed text and binary: read text first, then binary word = [] while True: ch = fin.read(1) if ch == b' ': break if ch != b'\n': # ignore newlines in front of words (some binary files have) word.append(ch) word = utils.to_unicode(b''.join(word), encoding=encoding, errors=unicode_errors) weights = fromstring(fin.read(binary_len), dtype=REAL) if word in self.wv.vocab: overlap_count += 1 self.wv.vectors[self.wv.vocab[word].index] = weights self.trainables.vectors_lockf[self.wv.vocab[word].index] = lockf # lock-factor: 0.0=no changes else: for line_no, line in enumerate(fin): parts = utils.to_unicode(line.rstrip(), encoding=encoding, errors=unicode_errors).split(" ") if len(parts) != vector_size + 1: raise ValueError("invalid vector on line %s (is this really the text format?)" % line_no) word, weights = parts[0], [REAL(x) for x in parts[1:]] if word in self.wv.vocab: overlap_count += 1 self.wv.vectors[self.wv.vocab[word].index] = weights self.trainables.vectors_lockf[self.wv.vocab[word].index] = lockf # lock-factor: 0.0=no changes logger.info("merged %d vectors into %s matrix from %s", overlap_count, self.wv.vectors.shape, fname) @deprecated("Method will be removed in 4.0.0, use self.wv.__getitem__() instead") def __getitem__(self, words): """Deprecated. Use `self.wv.__getitem__` instead. Refer to the documentation for :meth:`~gensim.models.keyedvectors.Word2VecKeyedVectors.__getitem__`. """ return self.wv.__getitem__(words) @deprecated("Method will be removed in 4.0.0, use self.wv.__contains__() instead") def __contains__(self, word): """Deprecated. Use `self.wv.__contains__` instead. Refer to the documentation for :meth:`~gensim.models.keyedvectors.Word2VecKeyedVectors.__contains__`. """ return self.wv.__contains__(word) def predict_output_word(self, context_words_list, topn=10): """Get the probability distribution of the center word given context words. Parameters ---------- context_words_list : list of str List of context words. topn : int, optional Return `topn` words and their probabilities. Returns ------- list of (str, float) `topn` length list of tuples of (word, probability). """ if not self.negative: raise RuntimeError( "We have currently only implemented predict_output_word for the negative sampling scheme, " "so you need to have run word2vec with negative > 0 for this to work." ) if not hasattr(self.wv, 'vectors') or not hasattr(self.trainables, 'syn1neg'): raise RuntimeError("Parameters required for predicting the output words not found.") word_vocabs = [self.wv.vocab[w] for w in context_words_list if w in self.wv.vocab] if not word_vocabs: warnings.warn("All the input context words are out-of-vocabulary for the current model.") return None word2_indices = [word.index for word in word_vocabs] l1 = np_sum(self.wv.vectors[word2_indices], axis=0) if word2_indices and self.cbow_mean: l1 /= len(word2_indices) # propagate hidden -> output and take softmax to get probabilities prob_values = exp(dot(l1, self.trainables.syn1neg.T)) prob_values /= sum(prob_values) top_indices = matutils.argsort(prob_values, topn=topn, reverse=True) # returning the most probable output words with their probabilities return [(self.wv.index2word[index1], prob_values[index1]) for index1 in top_indices] def init_sims(self, replace=False): """Deprecated. Use `self.wv.init_sims` instead. See :meth:`~gensim.models.keyedvectors.Word2VecKeyedVectors.init_sims`. """ if replace and hasattr(self.trainables, 'syn1'): del self.trainables.syn1 return self.wv.init_sims(replace) def reset_from(self, other_model): """Borrow shareable pre-built structures from `other_model` and reset hidden layer weights. Structures copied are: * Vocabulary * Index to word mapping * Cumulative frequency table (used for negative sampling) * Cached corpus length Useful when testing multiple models on the same corpus in parallel. Parameters ---------- other_model : :class:`~gensim.models.word2vec.Word2Vec` Another model to copy the internal structures from. """ self.wv.vocab = other_model.wv.vocab self.wv.index2word = other_model.wv.index2word self.vocabulary.cum_table = other_model.vocabulary.cum_table self.corpus_count = other_model.corpus_count self.trainables.reset_weights(self.hs, self.negative, self.wv) @staticmethod def log_accuracy(section): """Deprecated. Use `self.wv.log_accuracy` instead. See :meth:`~gensim.models.word2vec.Word2VecKeyedVectors.log_accuracy`. """ return Word2VecKeyedVectors.log_accuracy(section) @deprecated("Method will be removed in 4.0.0, use self.wv.evaluate_word_analogies() instead") def accuracy(self, questions, restrict_vocab=30000, most_similar=None, case_insensitive=True): """Deprecated. Use `self.wv.accuracy` instead. See :meth:`~gensim.models.word2vec.Word2VecKeyedVectors.accuracy`. """ most_similar = most_similar or Word2VecKeyedVectors.most_similar return self.wv.accuracy(questions, restrict_vocab, most_similar, case_insensitive) def __str__(self): """Human readable representation of the model's state. Returns ------- str Human readable representation of the model's state, including the vocabulary size, vector size and learning rate. """ return "%s(vocab=%s, size=%s, alpha=%s)" % ( self.__class__.__name__, len(self.wv.index2word), self.wv.vector_size, self.alpha ) def delete_temporary_training_data(self, replace_word_vectors_with_normalized=False): """Discard parameters that are used in training and scoring, to save memory. Warnings -------- Use only if you're sure you're done training a model. Parameters ---------- replace_word_vectors_with_normalized : bool, optional If True, forget the original (not normalized) word vectors and only keep the L2-normalized word vectors, to save even more memory. """ if replace_word_vectors_with_normalized: self.init_sims(replace=True) self._minimize_model() def save(self, *args, **kwargs): """Save the model. This saved model can be loaded again using :func:`~gensim.models.word2vec.Word2Vec.load`, which supports online training and getting vectors for vocabulary words. Parameters ---------- fname : str Path to the file. """ # don't bother storing the cached normalized vectors, recalculable table kwargs['ignore'] = kwargs.get('ignore', ['vectors_norm', 'cum_table']) super(Word2Vec, self).save(*args, **kwargs) def get_latest_training_loss(self): """Get current value of the training loss. Returns ------- float Current training loss. """ return self.running_training_loss @deprecated( "Method will be removed in 4.0.0, keep just_word_vectors = model.wv to retain just the KeyedVectors instance" ) def _minimize_model(self, save_syn1=False, save_syn1neg=False, save_vectors_lockf=False): if save_syn1 and save_syn1neg and save_vectors_lockf: return if hasattr(self.trainables, 'syn1') and not save_syn1: del self.trainables.syn1 if hasattr(self.trainables, 'syn1neg') and not save_syn1neg: del self.trainables.syn1neg if hasattr(self.trainables, 'vectors_lockf') and not save_vectors_lockf: del self.trainables.vectors_lockf self.model_trimmed_post_training = True @classmethod def load_word2vec_format( cls, fname, fvocab=None, binary=False, encoding='utf8', unicode_errors='strict', limit=None, datatype=REAL): """Deprecated. Use :meth:`gensim.models.KeyedVectors.load_word2vec_format` instead.""" raise DeprecationWarning("Deprecated. Use gensim.models.KeyedVectors.load_word2vec_format instead.") def save_word2vec_format(self, fname, fvocab=None, binary=False): """Deprecated. Use `model.wv.save_word2vec_format` instead. See :meth:`gensim.models.KeyedVectors.save_word2vec_format`. """ raise DeprecationWarning("Deprecated. Use model.wv.save_word2vec_format instead.") @classmethod def load(cls, *args, **kwargs): """Load a previously saved :class:`~gensim.models.word2vec.Word2Vec` model. See Also -------- :meth:`~gensim.models.word2vec.Word2Vec.save` Save model. Parameters ---------- fname : str Path to the saved file. Returns ------- :class:`~gensim.models.word2vec.Word2Vec` Loaded model. """ try: model = super(Word2Vec, cls).load(*args, **kwargs) # for backward compatibility for `max_final_vocab` feature if not hasattr(model, 'max_final_vocab'): model.max_final_vocab = None model.vocabulary.max_final_vocab = None return model except AttributeError: logger.info('Model saved using code from earlier Gensim Version. Re-loading old model in a compatible way.') from gensim.models.deprecated.word2vec import load_old_word2vec return load_old_word2vec(*args, **kwargs) class BrownCorpus(object): """Iterate over sentences from the `Brown corpus `_ (part of `NLTK data `_). """ def __init__(self, dirname): self.dirname = dirname def __iter__(self): for fname in os.listdir(self.dirname): fname = os.path.join(self.dirname, fname) if not os.path.isfile(fname): continue for line in utils.smart_open(fname): line = utils.to_unicode(line) # each file line is a single sentence in the Brown corpus # each token is WORD/POS_TAG token_tags = [t.split('/') for t in line.split() if len(t.split('/')) == 2] # ignore words with non-alphabetic tags like ",", "!" etc (punctuation, weird stuff) words = ["%s/%s" % (token.lower(), tag[:2]) for token, tag in token_tags if tag[:2].isalpha()] if not words: # don't bother sending out empty sentences continue yield words class Text8Corpus(object): """Iterate over sentences from the "text8" corpus, unzipped from http://mattmahoney.net/dc/text8.zip.""" def __init__(self, fname, max_sentence_length=MAX_WORDS_IN_BATCH): self.fname = fname self.max_sentence_length = max_sentence_length def __iter__(self): # the entire corpus is one gigantic line -- there are no sentence marks at all # so just split the sequence of tokens arbitrarily: 1 sentence = 1000 tokens sentence, rest = [], b'' with utils.smart_open(self.fname) as fin: while True: text = rest + fin.read(8192) # avoid loading the entire file (=1 line) into RAM if text == rest: # EOF words = utils.to_unicode(text).split() sentence.extend(words) # return the last chunk of words, too (may be shorter/longer) if sentence: yield sentence break last_token = text.rfind(b' ') # last token may have been split in two... keep for next iteration words, rest = (utils.to_unicode(text[:last_token]).split(), text[last_token:].strip()) if last_token >= 0 else ([], text) sentence.extend(words) while len(sentence) >= self.max_sentence_length: yield sentence[:self.max_sentence_length] sentence = sentence[self.max_sentence_length:] class LineSentence(object): """Iterate over a file that contains sentences: one line = one sentence. Words must be already preprocessed and separated by whitespace. """ def __init__(self, source, max_sentence_length=MAX_WORDS_IN_BATCH, limit=None): """ Parameters ---------- source : string or a file-like object Path to the file on disk, or an already-open file object (must support `seek(0)`). limit : int or None Clip the file to the first `limit` lines. Do no clipping if `limit is None` (the default). Examples -------- >>> from gensim.test.utils import datapath >>> sentences = LineSentence(datapath('lee_background.cor')) >>> for sentence in sentences: ... pass """ self.source = source self.max_sentence_length = max_sentence_length self.limit = limit def __iter__(self): """Iterate through the lines in the source.""" try: # Assume it is a file-like object and try treating it as such # Things that don't have seek will trigger an exception self.source.seek(0) for line in itertools.islice(self.source, self.limit): line = utils.to_unicode(line).split() i = 0 while i < len(line): yield line[i: i + self.max_sentence_length] i += self.max_sentence_length except AttributeError: # If it didn't work like a file, use it as a string filename with utils.smart_open(self.source) as fin: for line in itertools.islice(fin, self.limit): line = utils.to_unicode(line).split() i = 0 while i < len(line): yield line[i: i + self.max_sentence_length] i += self.max_sentence_length class PathLineSentences(object): """Like :class:`~gensim.models.word2vec.LineSentence`, but process all files in a directory in alphabetical order by filename. The directory must only contain files that can be read by :class:`gensim.models.word2vec.LineSentence`: .bz2, .gz, and text files. Any file not ending with .bz2 or .gz is assumed to be a text file. The format of files (either text, or compressed text files) in the path is one sentence = one line, with words already preprocessed and separated by whitespace. Warnings -------- Does **not recurse** into subdirectories. """ def __init__(self, source, max_sentence_length=MAX_WORDS_IN_BATCH, limit=None): """ Parameters ---------- source : str Path to the directory. limit : int or None Read only the first `limit` lines from each file. Read all if limit is None (the default). """ self.source = source self.max_sentence_length = max_sentence_length self.limit = limit if os.path.isfile(self.source): logger.debug('single file given as source, rather than a directory of files') logger.debug('consider using models.word2vec.LineSentence for a single file') self.input_files = [self.source] # force code compatibility with list of files elif os.path.isdir(self.source): self.source = os.path.join(self.source, '') # ensures os-specific slash at end of path logger.info('reading directory %s', self.source) self.input_files = os.listdir(self.source) self.input_files = [self.source + filename for filename in self.input_files] # make full paths self.input_files.sort() # makes sure it happens in filename order else: # not a file or a directory, then we can't do anything with it raise ValueError('input is neither a file nor a path') logger.info('files read into PathLineSentences:%s', '\n'.join(self.input_files)) def __iter__(self): """iterate through the files""" for file_name in self.input_files: logger.info('reading file %s', file_name) with utils.smart_open(file_name) as fin: for line in itertools.islice(fin, self.limit): line = utils.to_unicode(line).split() i = 0 while i < len(line): yield line[i:i + self.max_sentence_length] i += self.max_sentence_length class Word2VecVocab(utils.SaveLoad): """Vocabulary used by :class:`~gensim.models.word2vec.Word2Vec`.""" def __init__( self, max_vocab_size=None, min_count=5, sample=1e-3, sorted_vocab=True, null_word=0, max_final_vocab=None, ns_exponent=0.75): self.max_vocab_size = max_vocab_size self.min_count = min_count self.sample = sample self.sorted_vocab = sorted_vocab self.null_word = null_word self.cum_table = None # for negative sampling self.raw_vocab = None self.max_final_vocab = max_final_vocab self.ns_exponent = ns_exponent def scan_vocab(self, sentences, progress_per=10000, trim_rule=None): """Do an initial scan of all words appearing in sentences.""" logger.info("collecting all words and their counts") sentence_no = -1 total_words = 0 min_reduce = 1 vocab = defaultdict(int) checked_string_types = 0 for sentence_no, sentence in enumerate(sentences): if not checked_string_types: if isinstance(sentence, string_types): logger.warning( "Each 'sentences' item should be a list of words (usually unicode strings). " "First item here is instead plain %s.", type(sentence) ) checked_string_types += 1 if sentence_no % progress_per == 0: logger.info( "PROGRESS: at sentence #%i, processed %i words, keeping %i word types", sentence_no, total_words, len(vocab) ) for word in sentence: vocab[word] += 1 total_words += len(sentence) if self.max_vocab_size and len(vocab) > self.max_vocab_size: utils.prune_vocab(vocab, min_reduce, trim_rule=trim_rule) min_reduce += 1 logger.info( "collected %i word types from a corpus of %i raw words and %i sentences", len(vocab), total_words, sentence_no + 1 ) corpus_count = sentence_no + 1 self.raw_vocab = vocab return total_words, corpus_count def sort_vocab(self, wv): """Sort the vocabulary so the most frequent words have the lowest indexes.""" if len(wv.vectors): raise RuntimeError("cannot sort vocabulary after model weights already initialized.") wv.index2word.sort(key=lambda word: wv.vocab[word].count, reverse=True) for i, word in enumerate(wv.index2word): wv.vocab[word].index = i def prepare_vocab( self, hs, negative, wv, update=False, keep_raw_vocab=False, trim_rule=None, min_count=None, sample=None, dry_run=False): """Apply vocabulary settings for `min_count` (discarding less-frequent words) and `sample` (controlling the downsampling of more-frequent words). Calling with `dry_run=True` will only simulate the provided settings and report the size of the retained vocabulary, effective corpus length, and estimated memory requirements. Results are both printed via logging and returned as a dict. Delete the raw vocabulary after the scaling is done to free up RAM, unless `keep_raw_vocab` is set. """ min_count = min_count or self.min_count sample = sample or self.sample drop_total = drop_unique = 0 # set effective_min_count to min_count in case max_final_vocab isn't set self.effective_min_count = min_count # if max_final_vocab is specified instead of min_count # pick a min_count which satisfies max_final_vocab as well as possible if self.max_final_vocab is not None: sorted_vocab = sorted(self.raw_vocab.keys(), key=lambda word: self.raw_vocab[word], reverse=True) calc_min_count = 1 if self.max_final_vocab < len(sorted_vocab): calc_min_count = self.raw_vocab[sorted_vocab[self.max_final_vocab]] + 1 self.effective_min_count = max(calc_min_count, min_count) logger.info( "max_final_vocab=%d and min_count=%d resulted in calc_min_count=%d, effective_min_count=%d", self.max_final_vocab, min_count, calc_min_count, self.effective_min_count ) if not update: logger.info("Loading a fresh vocabulary") retain_total, retain_words = 0, [] # Discard words less-frequent than min_count if not dry_run: wv.index2word = [] # make stored settings match these applied settings self.min_count = min_count self.sample = sample wv.vocab = {} for word, v in iteritems(self.raw_vocab): if keep_vocab_item(word, v, self.effective_min_count, trim_rule=trim_rule): retain_words.append(word) retain_total += v if not dry_run: wv.vocab[word] = Vocab(count=v, index=len(wv.index2word)) wv.index2word.append(word) else: drop_unique += 1 drop_total += v original_unique_total = len(retain_words) + drop_unique retain_unique_pct = len(retain_words) * 100 / max(original_unique_total, 1) logger.info( "effective_min_count=%d retains %i unique words (%i%% of original %i, drops %i)", self.effective_min_count, len(retain_words), retain_unique_pct, original_unique_total, drop_unique ) original_total = retain_total + drop_total retain_pct = retain_total * 100 / max(original_total, 1) logger.info( "effective_min_count=%d leaves %i word corpus (%i%% of original %i, drops %i)", self.effective_min_count, retain_total, retain_pct, original_total, drop_total ) else: logger.info("Updating model with new vocabulary") new_total = pre_exist_total = 0 new_words = pre_exist_words = [] for word, v in iteritems(self.raw_vocab): if keep_vocab_item(word, v, self.effective_min_count, trim_rule=trim_rule): if word in wv.vocab: pre_exist_words.append(word) pre_exist_total += v if not dry_run: wv.vocab[word].count += v else: new_words.append(word) new_total += v if not dry_run: wv.vocab[word] = Vocab(count=v, index=len(wv.index2word)) wv.index2word.append(word) else: drop_unique += 1 drop_total += v original_unique_total = len(pre_exist_words) + len(new_words) + drop_unique pre_exist_unique_pct = len(pre_exist_words) * 100 / max(original_unique_total, 1) new_unique_pct = len(new_words) * 100 / max(original_unique_total, 1) logger.info( "New added %i unique words (%i%% of original %i) " "and increased the count of %i pre-existing words (%i%% of original %i)", len(new_words), new_unique_pct, original_unique_total, len(pre_exist_words), pre_exist_unique_pct, original_unique_total ) retain_words = new_words + pre_exist_words retain_total = new_total + pre_exist_total # Precalculate each vocabulary item's threshold for sampling if not sample: # no words downsampled threshold_count = retain_total elif sample < 1.0: # traditional meaning: set parameter as proportion of total threshold_count = sample * retain_total else: # new shorthand: sample >= 1 means downsample all words with higher count than sample threshold_count = int(sample * (3 + sqrt(5)) / 2) downsample_total, downsample_unique = 0, 0 for w in retain_words: v = self.raw_vocab[w] word_probability = (sqrt(v / threshold_count) + 1) * (threshold_count / v) if word_probability < 1.0: downsample_unique += 1 downsample_total += word_probability * v else: word_probability = 1.0 downsample_total += v if not dry_run: wv.vocab[w].sample_int = int(round(word_probability * 2**32)) if not dry_run and not keep_raw_vocab: logger.info("deleting the raw counts dictionary of %i items", len(self.raw_vocab)) self.raw_vocab = defaultdict(int) logger.info("sample=%g downsamples %i most-common words", sample, downsample_unique) logger.info( "downsampling leaves estimated %i word corpus (%.1f%% of prior %i)", downsample_total, downsample_total * 100.0 / max(retain_total, 1), retain_total ) # return from each step: words-affected, resulting-corpus-size, extra memory estimates report_values = { 'drop_unique': drop_unique, 'retain_total': retain_total, 'downsample_unique': downsample_unique, 'downsample_total': int(downsample_total), 'num_retained_words': len(retain_words) } if self.null_word: # create null pseudo-word for padding when using concatenative L1 (run-of-words) # this word is only ever input – never predicted – so count, huffman-point, etc doesn't matter self.add_null_word(wv) if self.sorted_vocab and not update: self.sort_vocab(wv) if hs: # add info about each word's Huffman encoding self.create_binary_tree(wv) if negative: # build the table for drawing random words (for negative sampling) self.make_cum_table(wv) return report_values def add_null_word(self, wv): word, v = '\0', Vocab(count=1, sample_int=0) v.index = len(wv.vocab) wv.index2word.append(word) wv.vocab[word] = v def create_binary_tree(self, wv): """Create a `binary Huffman tree `_ using stored vocabulary word counts. Frequent words will have shorter binary codes. Called internally from :meth:`~gensim.models.word2vec.Word2VecVocab.build_vocab`. """ logger.info("constructing a huffman tree from %i words", len(wv.vocab)) # build the huffman tree heap = list(itervalues(wv.vocab)) heapq.heapify(heap) for i in xrange(len(wv.vocab) - 1): min1, min2 = heapq.heappop(heap), heapq.heappop(heap) heapq.heappush( heap, Vocab(count=min1.count + min2.count, index=i + len(wv.vocab), left=min1, right=min2) ) # recurse over the tree, assigning a binary code to each vocabulary word if heap: max_depth, stack = 0, [(heap[0], [], [])] while stack: node, codes, points = stack.pop() if node.index < len(wv.vocab): # leaf node => store its path from the root node.code, node.point = codes, points max_depth = max(len(codes), max_depth) else: # inner node => continue recursion points = array(list(points) + [node.index - len(wv.vocab)], dtype=uint32) stack.append((node.left, array(list(codes) + [0], dtype=uint8), points)) stack.append((node.right, array(list(codes) + [1], dtype=uint8), points)) logger.info("built huffman tree with maximum node depth %i", max_depth) def make_cum_table(self, wv, domain=2**31 - 1): """Create a cumulative-distribution table using stored vocabulary word counts for drawing random words in the negative-sampling training routines. To draw a word index, choose a random integer up to the maximum value in the table (cum_table[-1]), then finding that integer's sorted insertion point (as if by `bisect_left` or `ndarray.searchsorted()`). That insertion point is the drawn index, coming up in proportion equal to the increment at that slot. Called internally from :meth:`~gensim.models.word2vec.Word2VecVocab.build_vocab`. """ vocab_size = len(wv.index2word) self.cum_table = zeros(vocab_size, dtype=uint32) # compute sum of all power (Z in paper) train_words_pow = 0.0 for word_index in xrange(vocab_size): train_words_pow += wv.vocab[wv.index2word[word_index]].count**self.ns_exponent cumulative = 0.0 for word_index in xrange(vocab_size): cumulative += wv.vocab[wv.index2word[word_index]].count**self.ns_exponent self.cum_table[word_index] = round(cumulative / train_words_pow * domain) if len(self.cum_table) > 0: assert self.cum_table[-1] == domain class Word2VecTrainables(utils.SaveLoad): """Represents the inner shallow neural network used to train :class:`~gensim.models.word2vec.Word2Vec`.""" def __init__(self, vector_size=100, seed=1, hashfxn=hash): self.hashfxn = hashfxn self.layer1_size = vector_size self.seed = seed def prepare_weights(self, hs, negative, wv, update=False, vocabulary=None): """Build tables and model weights based on final vocabulary settings.""" # set initial input/projection and hidden weights if not update: self.reset_weights(hs, negative, wv) else: self.update_weights(hs, negative, wv) def seeded_vector(self, seed_string, vector_size): """Get a random vector (but deterministic by seed_string).""" # Note: built-in hash() may vary by Python version or even (in Py3.x) per launch once = random.RandomState(self.hashfxn(seed_string) & 0xffffffff) return (once.rand(vector_size) - 0.5) / vector_size def reset_weights(self, hs, negative, wv): """Reset all projection weights to an initial (untrained) state, but keep the existing vocabulary.""" logger.info("resetting layer weights") wv.vectors = empty((len(wv.vocab), wv.vector_size), dtype=REAL) # randomize weights vector by vector, rather than materializing a huge random matrix in RAM at once for i in xrange(len(wv.vocab)): # construct deterministic seed from word AND seed argument wv.vectors[i] = self.seeded_vector(wv.index2word[i] + str(self.seed), wv.vector_size) if hs: self.syn1 = zeros((len(wv.vocab), self.layer1_size), dtype=REAL) if negative: self.syn1neg = zeros((len(wv.vocab), self.layer1_size), dtype=REAL) wv.vectors_norm = None self.vectors_lockf = ones(len(wv.vocab), dtype=REAL) # zeros suppress learning def update_weights(self, hs, negative, wv): """Copy all the existing weights, and reset the weights for the newly added vocabulary.""" logger.info("updating layer weights") gained_vocab = len(wv.vocab) - len(wv.vectors) newvectors = empty((gained_vocab, wv.vector_size), dtype=REAL) # randomize the remaining words for i in xrange(len(wv.vectors), len(wv.vocab)): # construct deterministic seed from word AND seed argument newvectors[i - len(wv.vectors)] = self.seeded_vector(wv.index2word[i] + str(self.seed), wv.vector_size) # Raise an error if an online update is run before initial training on a corpus if not len(wv.vectors): raise RuntimeError( "You cannot do an online vocabulary-update of a model which has no prior vocabulary. " "First build the vocabulary of your model with a corpus before doing an online update." ) wv.vectors = vstack([wv.vectors, newvectors]) if hs: self.syn1 = vstack([self.syn1, zeros((gained_vocab, self.layer1_size), dtype=REAL)]) if negative: self.syn1neg = vstack([self.syn1neg, zeros((gained_vocab, self.layer1_size), dtype=REAL)]) wv.vectors_norm = None # do not suppress learning for already learned words self.vectors_lockf = ones(len(wv.vocab), dtype=REAL) # zeros suppress learning # Example: ./word2vec.py -train data.txt -output vec.txt -size 200 -window 5 -sample 1e-4 \ # -negative 5 -hs 0 -binary 0 -cbow 1 -iter 3 if __name__ == "__main__": import argparse logging.basicConfig( format='%(asctime)s : %(threadName)s : %(levelname)s : %(message)s', level=logging.INFO ) logger.info("running %s", " ".join(sys.argv)) logger.info("using optimization %s", FAST_VERSION) # check and process cmdline input program = os.path.basename(sys.argv[0]) if len(sys.argv) < 2: print(globals()['__doc__'] % locals()) sys.exit(1) from gensim.models.word2vec import Word2Vec # noqa:F811 avoid referencing __main__ in pickle seterr(all='raise') # don't ignore numpy errors parser = argparse.ArgumentParser() parser.add_argument("-train", help="Use text data from file TRAIN to train the model", required=True) parser.add_argument("-output", help="Use file OUTPUT to save the resulting word vectors") parser.add_argument("-window", help="Set max skip length WINDOW between words; default is 5", type=int, default=5) parser.add_argument("-size", help="Set size of word vectors; default is 100", type=int, default=100) parser.add_argument( "-sample", help="Set threshold for occurrence of words. " "Those that appear with higher frequency in the training data will be randomly down-sampled;" " default is 1e-3, useful range is (0, 1e-5)", type=float, default=1e-3 ) parser.add_argument( "-hs", help="Use Hierarchical Softmax; default is 0 (not used)", type=int, default=0, choices=[0, 1] ) parser.add_argument( "-negative", help="Number of negative examples; default is 5, common values are 3 - 10 (0 = not used)", type=int, default=5 ) parser.add_argument("-threads", help="Use THREADS threads (default 12)", type=int, default=12) parser.add_argument("-iter", help="Run more training iterations (default 5)", type=int, default=5) parser.add_argument( "-min_count", help="This will discard words that appear less than MIN_COUNT times; default is 5", type=int, default=5 ) parser.add_argument( "-cbow", help="Use the continuous bag of words model; default is 1 (use 0 for skip-gram model)", type=int, default=1, choices=[0, 1] ) parser.add_argument( "-binary", help="Save the resulting vectors in binary mode; default is 0 (off)", type=int, default=0, choices=[0, 1] ) parser.add_argument("-accuracy", help="Use questions from file ACCURACY to evaluate the model") args = parser.parse_args() if args.cbow == 0: skipgram = 1 else: skipgram = 0 corpus = LineSentence(args.train) model = Word2Vec( corpus, size=args.size, min_count=args.min_count, workers=args.threads, window=args.window, sample=args.sample, sg=skipgram, hs=args.hs, negative=args.negative, cbow_mean=1, iter=args.iter ) if args.output: outfile = args.output model.wv.save_word2vec_format(outfile, binary=args.binary) else: outfile = args.train model.save(outfile + '.model') if args.binary == 1: model.wv.save_word2vec_format(outfile + '.model.bin', binary=True) else: model.wv.save_word2vec_format(outfile + '.model.txt', binary=False) if args.accuracy: model.accuracy(args.accuracy) logger.info("finished running %s", program)