-
Notifications
You must be signed in to change notification settings - Fork 0
/
EncoderDecoderModel.py
181 lines (149 loc) · 7.15 KB
/
EncoderDecoderModel.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
#!/usr/bin/env python
#-*- coding:utf-8 -*-
#!/usr/bin/python3
import sys
import math
import numpy as np
from argparse import ArgumentParser
from chainer import functions, optimizers
import util.generators as gens
from util.functions import trace, fill_batch
from util.model_file import ModelFile
from util.vocabulary import Vocabulary
import random
from util.chainer_cpu_wrapper import wrapper
#from util.chainer_gpu_wrapper import wrapper
class EncoderDecoderModel:
def __init__(self, parameter_dict):
self.parameter_dict = parameter_dict
self.source = parameter_dict["source"]
self.target = parameter_dict["target"]
self.test_source = parameter_dict["test_source"]
self.test_target = parameter_dict["test_target"]
self.reference_target = parameter_dict["reference_target"]
self.vocab = parameter_dict["vocab"]
self.embed = parameter_dict["embed"]
self.hidden = parameter_dict["hidden"]
self.epoch = parameter_dict["epoch"]
self.minibatch = parameter_dict["minibatch"]
self.generation_limit = parameter_dict["generation_limit"]
def make_model(self):
self.model = wrapper.make_model(
# encoder
weight_xi = functions.EmbedID(len(self.src_vocab), self.n_embed),
weight_ip = functions.Linear(self.n_embed, 4 * self.n_hidden),
weight_pp = functions.Linear(self.n_hidden, 4 * self.n_hidden),
# decoder
weight_pq = functions.Linear(self.n_hidden, 4 * self.n_hidden),
weight_qj = functions.Linear(self.n_hidden, self.n_embed),
weight_jy = functions.Linear(self.n_embed, len(self.trg_vocab)),
weight_yq = functions.EmbedID(len(self.trg_vocab), 4 * self.n_hidden),
weight_qq = functions.Linear(self.n_hidden, 4 * self.n_hidden),
)
def new(self, src_vocab, trg_vocab, n_embed, n_hidden, parameter_dict):
self.src_vocab = src_vocab
self.trg_vocab = trg_vocab
self.n_embed = n_embed
self.n_hidden = n_hidden
self.make_model()
return self
def save(self, filename):
with ModelFile(filename, 'w') as fp:
self.src_vocab.save(fp.get_file_pointer())
self.trg_vocab.save(fp.get_file_pointer())
fp.write(self.n_embed)
fp.write(self.n_hidden)
wrapper.begin_model_access(self.model)
fp.write_embed(self.model.weight_xi)
fp.write_linear(self.model.weight_ip)
fp.write_linear(self.model.weight_pp)
fp.write_linear(self.model.weight_pq)
fp.write_linear(self.model.weight_qj)
fp.write_linear(self.model.weight_jy)
fp.write_embed(self.model.weight_yq)
fp.write_linear(self.model.weight_qq)
wrapper.end_model_access(self.model)
def load(self, filename):
with ModelFile(filename) as fp:
self.src_vocab = Vocabulary.load(fp.get_file_pointer())
self.trg_vocab = Vocabulary.load(fp.get_file_pointer())
self.n_embed = int(fp.read())
self.n_hidden = int(fp.read())
self.make_model()
wrapper.begin_model_access(self.model)
fp.read_embed(self.model.weight_xi)
fp.read_linear(self.model.weight_ip)
fp.read_linear(self.model.weight_pp)
fp.read_linear(self.model.weight_pq)
fp.read_linear(self.model.weight_qj)
fp.read_linear(self.model.weight_jy)
fp.read_embed(self.model.weight_yq)
fp.read_linear(self.model.weight_qq)
wrapper.end_model_access(self.model)
return self
def init_optimizer(self):
self.__opt = optimizers.AdaGrad(lr=0.01)
self.__opt.setup(self.model)
def forward(self, is_training, src_batch, trg_batch = None, generation_limit = None):
pass
def train(self, src_batch, trg_batch):
self.__opt.zero_grads()
hyp_batch, accum_loss = self.forward(True, src_batch, trg_batch=trg_batch)
accum_loss.backward()
self.__opt.clip_grads(10)
self.__opt.update()
return hyp_batch
def predict(self, src_batch, generation_limit):
return self.forward(False, src_batch, generation_limit=generation_limit)
def train_model(self):
trace('making vocaburaries ...')
src_vocab = Vocabulary.new(gens.word_list(self.source), self.vocab)
trg_vocab = Vocabulary.new(gens.word_list(self.target), self.vocab)
trace('making model ...')
model = self.new(src_vocab, trg_vocab, self.embed, self.hidden, self.parameter_dict)
random_number = random.randint(0, self.minibatch)
for i_epoch in range(self.epoch):
trace('epoch %d/%d: ' % (i_epoch + 1, self.epoch))
trained = 0
gen1 = gens.word_list(self.source)
gen2 = gens.word_list(self.target)
gen3 = gens.batch(gens.sorted_parallel(gen1, gen2, 100 * self.minibatch), self.minibatch)
model.init_optimizer()
for src_batch, trg_batch in gen3:
src_batch = fill_batch(src_batch)
trg_batch = fill_batch(trg_batch)
K = len(src_batch)
hyp_batch = model.train(src_batch, trg_batch)
if trained == 0:
self.print_out(random_number, i_epoch, trained, src_batch, trg_batch, hyp_batch)
trained += K
trace('saving model ...')
model.save("ChainerMachineTranslation" + '.%03d' % (self.epoch + 1))
trace('finished.')
def test_model(self, model_name):
trace('loading model ...')
model = self.load(model_name)
trace('generating translation ...')
generated = 0
with open(self.test_target, 'w') as fp:
for src_batch in gens.batch(gens.word_list(self.test_source), self.minibatch):
src_batch = fill_batch(src_batch)
K = len(src_batch)
trace('sample %8d - %8d ...' % (generated + 1, generated + K))
hyp_batch = model.predict(src_batch, self.generation_limit)
source_cuont = 0
for hyp in hyp_batch:
hyp.append('</s>')
hyp = hyp[:hyp.index('</s>')]
# BLEUの結果を計算するため.
print("".join(src_batch[source_cuont]).replace("</s>", ""))
print(' '.join(hyp))
print(' '.join(hyp), file=fp)
source_cuont = source_cuont + 1
generated += K
trace('finished.')
def print_out(self, K, i_epoch, trained, src_batch, trg_batch, hyp_batch):
trace('epoch %3d/%3d, sample %8d' % (i_epoch + 1, self.epoch, trained + K + 1))
trace(' src = ' + ' '.join([x if x != '</s>' else '*' for x in src_batch[K]]))
trace(' trg = ' + ' '.join([x if x != '</s>' else '*' for x in trg_batch[K]]))
trace(' hyp = ' + ' '.join([x if x != '</s>' else '*' for x in hyp_batch[K]]))