zpwiki/pages/students/2016/darius_lindvai/dp2021/punc.py

247 lines
9.2 KiB
Python
Raw Normal View History

2020-06-05 12:49:42 +00:00
import numpy as np
2020-04-10 07:34:14 +00:00
import torch
import torch.autograd as autograd
import torch.nn as nn
import torch.optim as optim
2020-06-05 12:49:42 +00:00
from sklearn import metrics
from datetime import datetime
2020-04-10 07:34:14 +00:00
torch.manual_seed(1)
def argmax(vec):
# return the argmax as a python int
_, idx = torch.max(vec, 1)
return idx.item()
def prepare_sequence(seq, to_ix):
idxs = [to_ix[w] for w in seq]
return torch.tensor(idxs, dtype=torch.long)
# 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)))
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)
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))
# These two statements enforce the constraint that we never transfer
# to the start tag and we never transfer from the stop tag
self.transitions.data[tag_to_ix[START_TAG], :] = -10000
self.transitions.data[:, tag_to_ix[STOP_TAG]] = -10000
self.hidden = self.init_hidden()
def init_hidden(self):
return (torch.randn(2, 1, self.hidden_dim // 2),
torch.randn(2, 1, self.hidden_dim // 2))
def _forward_alg(self, feats):
2020-06-05 12:49:42 +00:00
# Do the forward algorithm to compute the partition function
2020-04-10 07:34:14 +00:00
init_alphas = torch.full((1, self.tagset_size), -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 = init_alphas
# Iterate through the sentence
for feat in feats:
alphas_t = [] # The forward tensors at this timestep
for next_tag in range(self.tagset_size):
2020-06-05 12:49:42 +00:00
# broadcast 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
2020-04-10 07:34:14 +00:00
trans_score = self.transitions[next_tag].view(1, -1)
2020-06-05 12:49:42 +00:00
# The ith entry of next_tag_var is the value for the
# edge (i -> next_tag) before we do log-sum-exp
2020-04-10 07:34:14 +00:00
next_tag_var = forward_var + trans_score + emit_score
2020-06-05 12:49:42 +00:00
# The forward variable for this tag is log-sum-exp of all the
# scores.
2020-04-10 07:34:14 +00:00
alphas_t.append(log_sum_exp(next_tag_var).view(1))
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 = torch.zeros(1)
tags = torch.cat([torch.tensor([self.tag_to_ix[START_TAG]], dtype=torch.long), 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.full((1, self.tagset_size), -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 = 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].view(1))
# 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
2020-06-05 12:49:42 +00:00
def forward(self, sentence): # dont confuse this with _forward_alg above.
2020-04-10 07:34:14 +00:00
# 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
START_TAG = "<START>"
STOP_TAG = "<STOP>"
EMBEDDING_DIM = 5
HIDDEN_DIM = 4
2020-06-05 12:49:42 +00:00
# Make up some training data
2020-05-05 18:37:06 +00:00
with open('/home/dlindvai/work/text.txt', 'r') as text2:
with open('/home/dlindvai/work/tags.txt', 'r') as tags2:
text1 = text2.read().splitlines()
tags1 = tags2.read().splitlines()
for line in text1:
text = line.replace("['", "").replace("']", "")
for line in tags1:
tags = line.replace("['", "").replace("']", "")
training_data = [( text.split() , tags.split() )]
#print(training_data)
2020-04-10 07:34:14 +00:00
word_to_ix = {}
for sentence, tags in training_data:
2020-06-05 12:49:42 +00:00
for word in sentence:
if word not in word_to_ix:
word_to_ix[word] = len(word_to_ix)
2020-04-10 07:34:14 +00:00
2020-06-05 12:49:42 +00:00
tag_to_ix = {"S": 0, "P": 1, "C": 2, "Q": 3, "N": 4, START_TAG: 5, STOP_TAG: 6}
2020-04-10 07:34:14 +00:00
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)
2020-06-05 12:49:42 +00:00
# Check predictions before training
2020-04-10 07:34:14 +00:00
with torch.no_grad():
2020-06-05 12:49:42 +00:00
precheck_sent = prepare_sequence(training_data[0][0], word_to_ix)
precheck_tags = torch.tensor([tag_to_ix[t] for t in training_data[0][1]], dtype=torch.long)
#print(model(precheck_sent))
2020-04-10 07:34:14 +00:00
2020-06-05 12:49:42 +00:00
# Print start time
start = datetime.now()
start_time = start.strftime("%H:%M:%S")
print("Start time = ", start_time)
2020-04-10 07:34:14 +00:00
2020-06-05 12:49:42 +00:00
for epoch in range(50):
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()
2020-04-10 07:34:14 +00:00
2020-06-05 12:49:42 +00:00
# Step 2. Get our inputs ready for the network, that is, turn them into Tensors of word indices.
sentence_in = prepare_sequence(sentence, word_to_ix)
targets = torch.tensor([tag_to_ix[t] for t in tags], dtype=torch.long)
2020-04-10 07:34:14 +00:00
2020-06-05 12:49:42 +00:00
# Step 3. Run our forward pass.
loss = model.neg_log_likelihood(sentence_in, targets)
2020-04-10 07:34:14 +00:00
2020-06-05 12:49:42 +00:00
# Step 4. Compute the loss, gradients, and update the parameters by calling optimizer.step()
loss.backward()
optimizer.step()
# Check predictions after training
2020-04-10 07:34:14 +00:00
with torch.no_grad():
2020-06-05 12:49:42 +00:00
precheck_sent = prepare_sequence(training_data[0][0], word_to_ix)
#print(model(precheck_sent))
# Error calculator
var = model(precheck_sent)
y_true = np.array(targets)
y_pred = np.array(var[1])
print(metrics.confusion_matrix(y_true, y_pred))
print(metrics.classification_report(y_true, y_pred, digits=3))
2020-05-05 18:37:06 +00:00
2020-06-05 12:49:42 +00:00
# Print finish time
finish = datetime.now()
finish_time = finish.strftime("%H:%M:%S")
print("Finish time = ", finish_time)