Open In App
Related Articles

Implement your own word2vec(skip-gram) model in Python

Improve Article
Save Article
Like Article

Prerequisite: Introduction to word2vec
Natural language processing (NLP) is a subfield of computer science and artificial intelligence concerned with the interactions between computers and human (natural) languages. 
In NLP techniques, we map the words and phrases (from vocabulary or corpus) to vectors of numbers to make the processing easier. These types of language modeling techniques are called word embeddings
In 2013, Google announced word2vec, a group of related models that are used to produce word embeddings.
Let’s implement our own skip-gram model (in Python) by deriving the backpropagation equations of our neural network.
In skip-gram architecture of word2vec, the input is the center word and the predictions are the context words. Consider an array of words W, if W(i) is the input (center word), then W(i-2), W(i-1), W(i+1), and W(i+2) are the context words if the sliding window size is 2. 


Let's define some variables :

V    Number of unique words in our corpus of text ( Vocabulary )
x    Input layer (One hot encoding of our input word ). 
N    Number of neurons in the hidden layer of neural network
W    Weights between input layer and hidden layer
W'   Weights between hidden layer and output layer
y    A softmax output layer having probabilities of every word in our vocabulary


Skip gram architecture

Skip gram architecture

Our neural network architecture is defined, now let’s do some math to derive the equations needed for gradient descent.

Forward Propagation:

Multiplying one hot encoding of the center word (denoted by x) with the first weight matrix W to get hidden layer matrix h (of size N x 1). 
h = W^T.x
( Vx1 )        ( NxV )   ( Vx1 ) 
Now we multiply the hidden layer vector h with second weight matrix W’ to get a new matrix u
u = W'^T.h
( Vx1 )        ( VxN )   ( Nx1 ) 
Note that we have to apply a softmax> to layer u to get our output layer y.
Let uj be jth neuron of layer u 
Let wj be the jth word in our vocabulary where j is any index 
Let Vwj be the jth column of matrix W’(column corresponding to a word wj)
u_{j} = V_{w_{ij}}^T.h$
( 1×1 )        ( 1xN )   ( Nx1 ) 
y = softmax(u) 
yj = softmax(uj
yj denotes the probability that wj is a context word 
P(w_{j}| w_{i}) = y_{j} = \frac{e^{u_{j}}}{\sum_{j'=1}^{v} e^{u_{j'}}}$
P(wj|wi) is the probability that wj is a context word, given wi is the input word.
Thus, our goal is to maximize P( wj* | wi ), where j* represents the indices of context words
Clearly, we want to maximize 
\prod_{c=1}^{C} \frac{e^{{u_{{j_c}^*}}}}{\sum_{j'=1}^{v} e^{u_{j'}}}$
where j*c are the vocabulary indexes of context words. Context words range from c = 1, 2, 3..C 
Let’s take a negative log-likelihood of this function to get our loss function, which we want to minimize 
E = - \log{ \left\{ {\prod_{c=1}^{C} \frac{e^{{u_{{j_c}^*}}}}{\sum_{j'=1}^{v} e^{u_{j'}}}} \right\} \,,\, E\, \, being\, our\, loss\, function$
Let t be the actual output vector from our training data, for a particular center word. It will have 1’s at the positions of context words and 0’s at all other places. tj*c are the 1’s of the context words. 
We can multiply u_{{j_c}^*}}}   with {t_{{j_c}^*}}
E = -log({\prod_{c=1}^{C} e^{ {u_{{j_c}^*}} }) + log(\sum_{j'=1}^{v} e^{u_{j'}})^C
Solving this equation we get our loss function as – 
E = -\sum_{c=1}^C {u_{{j_c}^*}}} } + C.log(\sum_{j'=1}^{v} e^{u_{j'}})

Back Propagation:

The parameters to be adjusted are in the matrices W and W’, hence we have to find the partial derivatives of our loss function with respect to W and W’ to apply the gradient descent algorithm. 
We have to find \frac{\partial E}{\partial W'}\, \, and\, \, \frac{\partial E}{\partial W}$
\frac{\partial E}{\partial w'_{ij}} = \frac{\partial E}{\partial u_j} . \frac{\partial u_j}{\partial w'_{ij}}$
\frac{\partial E}{\partial u_j}= - \sum_{c=1}^{C}{u_{{j_c}^*}} + C.\frac{1}{\sum_{j'=1}^{v} e^{u_{j'}}}.\frac{\partial}{\partial u_j}{\sum_{j=1}^{V}e^{u_j}}
\frac{\partial E}{\partial u_j}= - \sum_{c=1}^{C}1 + \sum_{j=1}^{V}y_j
\frac{\partial E}{\partial u_j}= y_j - t_j = e_j
\frac{\partial E}{\partial w'_{ij}} = e_j.\frac{\partial u_j}{\partial w'_{ij}} = e_j.\frac{\partial w'_{ij}*h_i}{\partial w'_{ij}}$
\frac{\partial E}{\partial w'_{ij}} = e_j.h_i
Now, Finding \frac{\partial E}{\partial w_{ij}}
\frac{\partial E}{\partial w_{ij}} = \frac{\partial E}{\partial u_j} . \frac{\partial u_j}{\partial w_{ij}}$
\frac{\partial E}{\partial w_{ij}} = \frac{\partial E}{\partial u_j} . \frac{\partial u_j}{\partial h_i} . \frac{\partial h_i}{\partial w_{ij}}$
\frac{\partial E}{\partial w_{ij}} = e_j . w'_{ij} . \frac{\partial w_{ij}*x_i}{\partial w_{ij}}$
\frac{\partial E}{\partial w_{ij}} = e_j . w'_{ij} . x_i$
Below is the implementation : 


import numpy as np
import string
from nltk.corpus import stopwords
def softmax(x):
    """Compute softmax values for each sets of scores in x."""
    e_x = np.exp(x - np.max(x))
    return e_x / e_x.sum()
class word2vec(object):
    def __init__(self):
        self.N = 10
        self.X_train = []
        self.y_train = []
        self.window_size = 2
        self.alpha = 0.001
        self.words = []
        self.word_index = {}
    def initialize(self,V,data):
        self.V = V
        self.W = np.random.uniform(-0.8, 0.8, (self.V, self.N))
        self.W1 = np.random.uniform(-0.8, 0.8, (self.N, self.V))
        self.words = data
        for i in range(len(data)):
            self.word_index[data[i]] = i
    def feed_forward(self,X):
        self.h =,X).reshape(self.N,1)
        self.u =,self.h)
        self.y = softmax(self.u) 
        return self.y
    def backpropagate(self,x,t):
        e = self.y - np.asarray(t).reshape(self.V,1)
        # e.shape is V x 1
        dLdW1 =,e.T)
        X = np.array(x).reshape(self.V,1)
        dLdW =,,e).T)
        self.W1 = self.W1 - self.alpha*dLdW1
        self.W = self.W - self.alpha*dLdW
    def train(self,epochs):
        for x in range(1,epochs):       
            self.loss = 0
            for j in range(len(self.X_train)):
                C = 0
                for m in range(self.V):
                        self.loss += -1*self.u[m][0]
                        C += 1
                self.loss += C*np.log(np.sum(np.exp(self.u)))
            print("epoch ",x, " loss = ",self.loss)
            self.alpha *= 1/( (1+self.alpha*x) )
    def predict(self,word,number_of_predictions):
        if word in self.words:
            index = self.word_index[word]
            X = [0 for i in range(self.V)]
            X[index] = 1
            prediction = self.feed_forward(X)
            output = {}
            for i in range(self.V):
                output[prediction[i][0]] = i
            top_context_words = []
            for k in sorted(output,reverse=True):
            return top_context_words
            print("Word not found in dictionary")


def preprocessing(corpus):
    stop_words = set(stopwords.words('english'))   
    training_data = []
    sentences = corpus.split(".")
    for i in range(len(sentences)):
        sentences[i] = sentences[i].strip()
        sentence = sentences[i].split()
        x = [word.strip(string.punctuation) for word in sentence
                                     if word not in stop_words]
        x = [word.lower() for word in x]
    return training_data
def prepare_data_for_training(sentences,w2v):
    data = {}
    for sentence in sentences:
        for word in sentence:
            if word not in data:
                data[word] = 1
                data[word] += 1
    V = len(data)
    data = sorted(list(data.keys()))
    vocab = {}
    for i in range(len(data)):
        vocab[data[i]] = i
    #for i in range(len(words)):
    for sentence in sentences:
        for i in range(len(sentence)):
            center_word = [0 for x in range(V)]
            center_word[vocab[sentence[i]]] = 1
            context = [0 for x in range(V)]
            for j in range(i-w2v.window_size,i+w2v.window_size):
                if i!=j and j>=0 and j<len(sentence):
                    context[vocab[sentence[j]]] += 1
    return w2v.X_train,w2v.y_train  


corpus = ""
corpus += "The earth revolves around the sun. The moon revolves around the earth"
epochs = 1000
training_data = preprocessing(corpus)
w2v = word2vec()



Whether you're preparing for your first job interview or aiming to upskill in this ever-evolving tech landscape, GeeksforGeeks Courses are your key to success. We provide top-quality content at affordable prices, all geared towards accelerating your growth in a time-bound manner. Join the millions we've already empowered, and we're here to do the same for you. Don't miss out - check it out now!

Last Updated : 04 Aug, 2022
Like Article
Save Article
Similar Reads
Complete Tutorials