Bi-LSTM_CRF_4_tagger.py

# -*- coding: utf-8 -*-
# author: Sendong (Stan) Zhao

import torch
import torch.nn as nn
import torch.optim as optim
from torch.autograd import Variable
import datetime
import os

os.environ["OMP_NUM_THREADS"] = "6"

torch.manual_seed(1)

start = datetime.datetime.now()


def to_scalar(var):
    # returns a python float
    return var.view(-1).data.tolist()[0]


def argmax(vec):
    # return the argmax as a python int
    _, idx = torch.max(vec, 1)
    return to_scalar(idx)


def prepare_sequence(seq, to_ix):
    idxs = []
    for i, w in enumerate(seq):
        if not to_ix.has_key(w):
            idxs.append(0)
        else:
            idxs.append(to_ix[w]) 
    #idxs = [to_ix[w] for w in seq] 
    tensor = torch.LongTensor(idxs)
    return Variable(tensor)

# Compute log sum exp in a numerically stable way for the forward algorithm
def log_sum_exp(vec):
    max_score = vec[0, argmax(vec)]
    max_score_broadcast = max_score.view(1, -1).expand(1, vec.size()[1])
    return max_score + \
        torch.log(torch.sum(torch.exp(vec - max_score_broadcast)))

#####################################################################
# Create model


class BiLSTM_CRF(nn.Module):

    def __init__(self, vocab_size, tag_to_ix, embedding_dim, hidden_dim):
        super(BiLSTM_CRF, self).__init__()
        self.embedding_dim = embedding_dim
        self.hidden_dim = hidden_dim
        self.vocab_size = vocab_size
        self.tag_to_ix = tag_to_ix
        self.tagset_size = len(tag_to_ix)

        self.word_embeds = nn.Embedding(vocab_size, embedding_dim, padding_idx = 0)
        self.lstm = nn.LSTM(embedding_dim, hidden_dim // 2,
                            num_layers=1, bidirectional=True)

        # Maps the output of the LSTM into tag space.
        self.hidden2tag = nn.Linear(hidden_dim, self.tagset_size)

        # Matrix of transition parameters.  Entry i,j is the score of
        # transitioning *to* i *from* j.
        self.transitions = nn.Parameter(
            torch.randn(self.tagset_size, self.tagset_size))

        self.hidden = self.init_hidden()

    def init_hidden(self):
        return (Variable(torch.randn(2, 1, self.hidden_dim // 2)),
                Variable(torch.randn(2, 1, self.hidden_dim // 2)))

    def _forward_alg(self, feats):
        # Do the forward algorithm to compute the partition function
        init_alphas = torch.Tensor(1, self.tagset_size).fill_(-10000.)
        # START_TAG has all of the score.
        init_alphas[0][self.tag_to_ix[START_TAG]] = 0.

        # Wrap in a variable so that we will get automatic backprop
        forward_var = Variable(init_alphas)

        # Iterate through the sentence
        for feat in feats:
            alphas_t = []  # The forward variables at this timestep
            for next_tag in range(self.tagset_size):
                # broadcaist the emission score: it is the same regardless of
                # the previous tag
                emit_score = feat[next_tag].view(
                    1, -1).expand(1, self.tagset_size)
                # the ith entry of trans_score is the score of transitioning to
                # next_tag from i
                trans_score = self.transitions[next_tag].view(1, -1)
                # The ith entry of next_tag_var is the value for the
                # edge (i -> next_tag) before we do log-sum-exp
                next_tag_var = forward_var + trans_score + emit_score
                # The forward variable for this tag is log-sum-exp of all the
                # scores.
                alphas_t.append(log_sum_exp(next_tag_var))
            forward_var = torch.cat(alphas_t).view(1, -1)
        terminal_var = forward_var + self.transitions[self.tag_to_ix[STOP_TAG]]
        alpha = log_sum_exp(terminal_var)
        return alpha

    def _get_lstm_features(self, sentence):
        self.hidden = self.init_hidden()
        embeds = self.word_embeds(sentence).view(len(sentence), 1, -1)
        lstm_out, self.hidden = self.lstm(embeds, self.hidden)
        lstm_out = lstm_out.view(len(sentence), self.hidden_dim)
        lstm_feats = self.hidden2tag(lstm_out)
        return lstm_feats

    def _score_sentence(self, feats, tags):
        # Gives the score of a provided tag sequence
        score = Variable(torch.Tensor([0]))
        tags = torch.cat([torch.LongTensor([self.tag_to_ix[START_TAG]]), tags])
        for i, feat in enumerate(feats):
            score = score + \
                self.transitions[tags[i + 1], tags[i]] + feat[tags[i + 1]]
        score = score + self.transitions[self.tag_to_ix[STOP_TAG], tags[-1]]
        return score

    def _viterbi_decode(self, feats):
        backpointers = []

        # Initialize the viterbi variables in log space
        init_vvars = torch.Tensor(1, self.tagset_size).fill_(-10000.)
        init_vvars[0][self.tag_to_ix[START_TAG]] = 0

        # forward_var at step i holds the viterbi variables for step i-1
        forward_var = Variable(init_vvars)
        for feat in feats:
            bptrs_t = []  # holds the backpointers for this step
            viterbivars_t = []  # holds the viterbi variables for this step

            for next_tag in range(self.tagset_size):
                # next_tag_var[i] holds the viterbi variable for tag i at the
                # previous step, plus the score of transitioning
                # from tag i to next_tag.
                # We don't include the emission scores here because the max
                # does not depend on them (we add them in below)
                next_tag_var = forward_var + self.transitions[next_tag]
                best_tag_id = argmax(next_tag_var)
                bptrs_t.append(best_tag_id)
                viterbivars_t.append(next_tag_var[0][best_tag_id])
            # Now add in the emission scores, and assign forward_var to the set
            # of viterbi variables we just computed
            forward_var = (torch.cat(viterbivars_t) + feat).view(1, -1)
            backpointers.append(bptrs_t)

        # Transition to STOP_TAG
        terminal_var = forward_var + self.transitions[self.tag_to_ix[STOP_TAG]]
        best_tag_id = argmax(terminal_var)
        path_score = terminal_var[0][best_tag_id]

        # Follow the back pointers to decode the best path.
        best_path = [best_tag_id]
        for bptrs_t in reversed(backpointers):
            best_tag_id = bptrs_t[best_tag_id]
            best_path.append(best_tag_id)
        # Pop off the start tag (we dont want to return that to the caller)
        start = best_path.pop()
        assert start == self.tag_to_ix[START_TAG]  # Sanity check
        best_path.reverse()
        return path_score, best_path

    def neg_log_likelihood(self, sentence, tags):
        feats = self._get_lstm_features(sentence)
        forward_score = self._forward_alg(feats)
        gold_score = self._score_sentence(feats, tags)
        return forward_score - gold_score

    def forward(self, sentence):  # dont confuse this with _forward_alg above.
        # Get the emission scores from the BiLSTM
        lstm_feats = self._get_lstm_features(sentence)

        # Find the best path, given the features.
        score, tag_seq = self._viterbi_decode(lstm_feats)
        return score, tag_seq

#####################################################################
# Run training


START_TAG = "<START>"
STOP_TAG = "<STOP>"
EMBEDDING_DIM = 100
HIDDEN_DIM = 100

#Read data from file
def read_corpus(filename):
    data = []
    data_string_list = list(open(filename))

    print len(data_string_list)

    element_size = 0
    X = list()
    Y = list()

    for data_string in data_string_list:
        words = data_string.strip().split()
        if len(words) is 0:
            data.append((X,Y))
            X = list()
            Y = list()
        else:
            if element_size is 0:
                element_size = len(words)
            X.append(words[0])
            Y.append(words[-1])
    if len(X)>0:
        data.append((X,Y))

    return data

training_data = read_corpus('test')

word_to_ix = {}
tag_to_ix = {START_TAG: 0, STOP_TAG: 1}

for sent, tags in training_data:
    for word in sent:
         if word not in word_to_ix:
             word_to_ix[word] = len(word_to_ix)
    
    for tag in tags:
        if tag not in tag_to_ix:
            tag_to_ix[tag] = len(tag_to_ix)

print (len(tag_to_ix))

model = BiLSTM_CRF(len(word_to_ix), tag_to_ix, EMBEDDING_DIM, HIDDEN_DIM)
optimizer = optim.SGD(model.parameters(), lr=0.01, weight_decay=1e-4)

# Check predictions before training
precheck_sent = prepare_sequence(training_data[0][0], word_to_ix)
precheck_tags = torch.LongTensor([tag_to_ix[t] for t in training_data[0][1]])
print(model(precheck_sent))

# Make sure prepare_sequence from earlier in the LSTM section is loaded
for epoch in range(
        300):  # again, normally you would NOT do 300 epochs, it is toy data
    for sentence, tags in training_data:
        # Step 1. Remember that Pytorch accumulates gradients.
        # We need to clear them out before each instance
        model.zero_grad()

        # Step 2. Get our inputs ready for the network, that is,
        # turn them into Variables of word indices.
        sentence_in = prepare_sequence(sentence, word_to_ix)
        targets = torch.LongTensor([tag_to_ix[t] for t in tags])

        # Step 3. Run our forward pass.
        neg_log_likelihood = model.neg_log_likelihood(sentence_in, targets)

        # Step 4. Compute the loss, gradients, and update the parameters by
        # calling optimizer.step()
        neg_log_likelihood.backward()
        optimizer.step()

# Check predictions after training
precheck_sent = prepare_sequence(training_data[0][0], word_to_ix)
print(model(precheck_sent))

# run testing
#torch.save(model.state_dict(), 'Bi-LSTM_CRF_4_tagger.pkl')
testing_data = read_corpus('test')
#model.load_state_dict(torch.load('Bi-LSTM_CRF_4_tagger.pkl'))

total_count = 0
correct_count = 0
wrong_count = 0

for sentence, tags in testing_data:
    
    sentence_in = prepare_sequence(sentence, word_to_ix)
    targets = prepare_sequence(tags, tag_to_ix)
    #print (targets)
    tag_scores, idx = model(sentence_in)
    #print (idx)
    #print (tag_scores)
    
    for t in range(len(targets)):
        total_count += 1
        index = idx[t]
        
        if targets[t].data[0] == index:
            correct_count += 1
        else:
            wrong_count += 1

print('Correct: %d' % correct_count)
print('Wrong: %d' % wrong_count)
print('Total: %d' % total_count)
print('Performance: %f' % (float(correct_count)/float(total_count)))


end = datetime.datetime.now()

print (end - start)