-
Notifications
You must be signed in to change notification settings - Fork 13
/
pure_GFAW.py
182 lines (152 loc) · 8.22 KB
/
pure_GFAW.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
# ======================= import packages from Rl model =======================
from __future__ import absolute_import
from __future__ import division
from tqdm import tqdm
import json
import time
import os
import logging
import numpy as np
import tensorflow as tf
# Export PYTHONPATH so that 'rl_code' folder can be regarded as a package
import sys
from rl_code.model.trainer import Trainer
from rl_code.model.agent import Agent
from rl_code.options import read_options
from rl_code.model.environment import env
import codecs
from collections import defaultdict
import gc
import resource
import sys
from rl_code.model.baseline import ReactiveBaseline
from rl_code.model.nell_eval import nell_eval
from scipy.misc import logsumexp as lse
from pprint import pprint
# ======================= import packages from PCNN model =======================
import nrekit
import numpy as np
import tensorflow as tf
import sys
import json
import time
# ======================= initialize global parameters =======================
# tf and config settings
# os.environ["CUDA_VISIBLE_DEVICES"] = "1,4"
config = tf.ConfigProto()
config.gpu_options.allow_growth = True
config.log_device_placement = False
config.allow_soft_placement = True
# read command line options
options = read_options()
tf.set_random_seed(options['random_seed'])
np.random.seed(options['random_seed'])
# ======================= initialize GFAW parameters =======================
logger = logging.getLogger()
logging.basicConfig(stream=sys.stdout, level=logging.DEBUG)
# Set logging
logger.setLevel(logging.INFO)
fmt = logging.Formatter('%(asctime)s: [ %(message)s ]',
'%m/%d/%Y %I:%M:%S %p')
console = logging.StreamHandler()
console.setFormatter(fmt)
logger.addHandler(console)
logfile = logging.FileHandler(options['log_file_name'], 'w')
logfile.setFormatter(fmt)
logger.addHandler(logfile)
# read the vocab files, it will be used by many classes hence global scope
logger.info('reading vocab files...')
options['relation_vocab'] = json.load(open(options['vocab_dir'] + '/relation_vocab.json'))
options['entity_vocab'] = json.load(open(options['vocab_dir'] + '/entity_vocab.json'))
logger.info('Reading mid to name map')
mid_to_word = {}
logger.info('Done..')
logger.info('Total number of entities {}'.format(len(options['entity_vocab'])))
logger.info('Total number of relations {}'.format(len(options['relation_vocab'])))
# ======================= import code from PCNN model =======================
dataset_dir = os.path.join(options['pcnn_dataset_base'], options['pcnn_dataset_name'])
if not os.path.isdir(dataset_dir):
raise Exception("[ERROR] Dataset dir %s doesn't exist!" % (dataset_dir))
# The first 3 parameters are train / test data file name, word embedding file name and relation-id mapping file name respectively.
train_loader = nrekit.data_loader.json_file_data_loader(os.path.join(dataset_dir, 'train.json'),
os.path.join(dataset_dir, 'word_vec.json'),
os.path.join(dataset_dir, 'rel2id.json'),
mode=nrekit.data_loader.json_file_data_loader.MODE_RELFACT_BAG,
shuffle=True, batch_size=options['pcnn_batch_size'], case_sensitive=False,
reprocess=False)
test_loader = nrekit.data_loader.json_file_data_loader(os.path.join(dataset_dir, 'test.json'),
os.path.join(dataset_dir, 'word_vec.json'),
os.path.join(dataset_dir, 'rel2id.json'),
mode=nrekit.data_loader.json_file_data_loader.MODE_ENTPAIR_BAG,
shuffle=False, batch_size=options['pcnn_batch_size'], case_sensitive=False,
reprocess=False)
framework = nrekit.framework.re_framework(train_loader, test_loader)
class pcnn_model(nrekit.framework.re_model):
def __init__(self, train_data_loader, batch_size, max_length=120):
nrekit.framework.re_model.__init__(self, train_data_loader, batch_size, max_length=max_length)
self.mask = tf.placeholder(dtype=tf.int32, shape=[None, max_length], name="mask")
# Embedding
x = nrekit.network.embedding.word_position_embedding(self.word, self.word_vec_mat, self.pos1, self.pos2)
# Encoder
x_train = nrekit.network.encoder.pcnn(x, self.mask, keep_prob=0.5)
x_test = nrekit.network.encoder.pcnn(x, self.mask, keep_prob=1.0)
# Selector
self._train_logit, train_repre = nrekit.network.selector.bag_attention(x_train, self.scope,
self.ins_label,
self.rel_tot, True,
keep_prob=0.5)
self._test_logit, test_repre = nrekit.network.selector.bag_attention(x_test, self.scope,
self.ins_label,
self.rel_tot, False,
keep_prob=1.0)
# Classifier
self._loss = nrekit.network.classifier.softmax_cross_entropy(self._train_logit, self.label,
self.rel_tot,
weights_table=self.get_weights())
def loss(self):
return self._loss
def train_logit(self):
return self._train_logit
def test_logit(self):
return self._test_logit
def get_weights(self):
with tf.variable_scope("weights_table", reuse=tf.AUTO_REUSE):
print("Calculating weights_table...")
_weights_table = np.zeros((self.rel_tot), dtype=np.float32)
for i in range(len(self.train_data_loader.data_rel)):
_weights_table[self.train_data_loader.data_rel[i]] += 1.0
_weights_table = 1 / (_weights_table ** 0.05)
weights_table = tf.get_variable(name='weights_table', dtype=tf.float32, trainable=False,
initializer=_weights_table)
print("Finish calculating")
return weights_table
# TODO: Pretrain GFAW/PCNN if there doesn't exist
# ======================= Pretrain GFAW =======================
if not options['load_model']:
trainer = Trainer(options)
with tf.Session(config=config) as sess:
sess.run(trainer.initialize())
trainer.initialize_pretrained_embeddings(sess=sess)
# trainer.test_environment = trainer.test_test_environment # use test to show result
trainer.train(sess)
tf.reset_default_graph()
# Pretrained Model Test
trainer = Trainer(options)
save_path = os.path.join(options['model_dir'], 'model.ckpt') # trainer.save_path
path_logger_file = trainer.path_logger_file
output_dir = trainer.output_dir
with tf.Session(config=config) as sess:
trainer.initialize(restore=save_path, sess=sess)
trainer.test_rollouts = 100
test_set_name = 'test'
if not os.path.isdir(path_logger_file + "/" + "test_beam_" + test_set_name):
os.mkdir(path_logger_file + "/" + "test_beam_" + test_set_name)
trainer.path_logger_file_ = path_logger_file + "/" + "test_beam_" + test_set_name + "/paths"
with open(output_dir + '/scores.txt', 'a') as score_file:
score_file.write(test_set_name + "(beam) scores with best model from " + save_path + "\n")
# trainer.test_environment = trainer.test_test_environment # test_environment = dev_test_environment
trainer.test_environment = env(trainer.params, test_set_name)
# trainer.test_environment.test_rollouts = 100
trainer.train_environment.grapher.array_store = np.load(file=str(options['model_dir'] + 'new_graph.npy'))
trainer.test(sess, beam=True, print_paths=True, save_model=False)
tf.reset_default_graph()