-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathtrain_sent_senti_cls_rnn.py
203 lines (177 loc) · 8.54 KB
/
train_sent_senti_cls_rnn.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
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
import torch
import numpy as np
import os
import time
import sys
import pdb
import traceback
from bdb import BdbQuit
import json
import random
import tqdm
from copy import deepcopy
from collections import defaultdict
from opts import parse_opt
from models.sent_senti_cls import SentenceSentimentClassifier
from dataloader import get_senti_sents_dataloader
random.seed(100)
resume = ''
def clip_gradient(optimizer, grad_clip):
for group in optimizer.param_groups:
for param in group['params']:
if param.grad is not None:
param.grad.data.clamp_(-grad_clip, grad_clip)
def train():
dataset_name = opt.dataset_name
corpus_type = opt.corpus_type
idx2word = json.load(open(os.path.join(opt.captions_dir, dataset_name, corpus_type, 'idx2word.json'), 'r'))
senti_captions = json.load(open(os.path.join(opt.captions_dir, dataset_name, corpus_type, 'senti_captions.json'), 'r'))
model = SentenceSentimentClassifier(idx2word, opt.sentiment_categories, opt.settings)
model.to(opt.device)
lr = 4e-4
optimizer, criterion = model.get_optim_and_crit(lr)
if resume:
print("====> loading checkpoint '{}'".format(resume))
chkpoint = torch.load(resume, map_location=lambda s, l: s)
assert opt.settings == chkpoint['settings'], \
'opt.settings and resume model settings are different'
assert idx2word == chkpoint['idx2word'], \
'idx2word and resume model idx2word are different'
assert opt.sentiment_categories == chkpoint['sentiment_categories'], \
'sentiment_categories and resume model sentiment_categories are different'
assert dataset_name == chkpoint['dataset_name'], \
'dataset_name and resume model dataset_name are different'
assert corpus_type == chkpoint['corpus_type'], \
'corpus_type and resume model corpus_type are different'
model.load_state_dict(chkpoint['model'])
optimizer.load_state_dict(chkpoint['optimizer'])
lr = optimizer.param_groups[0]['lr']
print("====> loaded checkpoint '{}', epoch: {}"
.format(resume, chkpoint['epoch']))
word2idx = {}
for i, w in enumerate(idx2word):
word2idx[w] = i
senti_label2idx = {}
for i, w in enumerate(opt.sentiment_categories):
senti_label2idx[w] = i
print('====> process senti_corpus begin')
for senti in senti_captions:
senti_captions[senti] = [c[0] for c in senti_captions[senti]]
random.shuffle(senti_captions[senti])
tmp_senti_captions = {'train': {}, 'val': {}}
tmp_senti_captions['train']['neutral'] = deepcopy(senti_captions['neutral'][5000:])
tmp_senti_captions['val']['neutral'] = deepcopy(senti_captions['neutral'][:5000])
tmp_senti_captions['train']['positive'] = deepcopy(senti_captions['positive'][1000:])
tmp_senti_captions['val']['positive'] = deepcopy(senti_captions['positive'][:1000])
tmp_senti_captions['train']['negative'] = deepcopy(senti_captions['negative'][1000:])
tmp_senti_captions['val']['negative'] = deepcopy(senti_captions['negative'][:1000])
tmp_senti_captions['train']['positive'] = tmp_senti_captions['train']['positive'] * int(len(tmp_senti_captions['train']['neutral']) / len(tmp_senti_captions['train']['positive']))
tmp_senti_captions['train']['negative'] = tmp_senti_captions['train']['negative'] * int(len(tmp_senti_captions['train']['neutral']) / len(tmp_senti_captions['train']['negative']))
senti_captions = tmp_senti_captions
train_set = []
val_set = {}
for senti in opt.sentiment_categories:
print('convert %s corpus to index' % senti)
senti_id = senti_label2idx[senti]
for cap in tqdm.tqdm(senti_captions['train'][senti]):
tmp = [word2idx.get(w, None) or word2idx['<UNK>'] for w in cap] + [word2idx['<EOS>']]
train_set.append([senti_id, tmp])
val_set[senti] = []
for cap in tqdm.tqdm(senti_captions['val'][senti]):
tmp = [word2idx.get(w, None) or word2idx['<UNK>'] for w in cap] + [word2idx['<EOS>']]
val_set[senti].append([senti_id, tmp])
random.shuffle(train_set)
print('====> process senti_corpus end')
train_data = get_senti_sents_dataloader(train_set, word2idx['<PAD>'], opt.max_seq_len)
val_data = {}
for senti in val_set:
val_data[senti] = get_senti_sents_dataloader(val_set[senti], word2idx['<PAD>'], opt.max_seq_len, shuffle=False)
checkpoint = os.path.join(opt.checkpoint, 'sent_senti_cls', dataset_name, corpus_type)
if not os.path.exists(checkpoint):
os.makedirs(checkpoint)
result_dir = os.path.join(opt.result_dir, 'sent_senti_cls', dataset_name, corpus_type)
if not os.path.exists(result_dir):
os.makedirs(result_dir)
previous_acc_rate = None
for epoch in range(30):
print('--------------------epoch: %d' % epoch)
model.train()
train_loss = 0.0
for sentis, (caps_tensor, lengths) in tqdm.tqdm(train_data):
sentis = sentis.to(opt.device)
caps_tensor = caps_tensor.to(opt.device)
pred, _ = model(caps_tensor, lengths)
loss = criterion(pred, sentis)
train_loss += loss.item()
optimizer.zero_grad()
loss.backward()
clip_gradient(optimizer, opt.grad_clip)
optimizer.step()
train_loss /= len(train_data)
model.eval()
all_num = defaultdict(int)
senti_num = {}
test_case = defaultdict(list)
for senti in opt.sentiment_categories:
senti_num[senti] = defaultdict(int)
with torch.no_grad():
for senti, data in val_data.items():
for sentis, (caps_tensor, lengths) in tqdm.tqdm(data):
sentis = sentis.to(opt.device)
caps_tensor = caps_tensor.to(opt.device)
rest, rest_w, att_weights = model.sample(caps_tensor, lengths)
rest = torch.LongTensor(np.array(rest)).to(opt.device)
total_num = int(sentis.size(0))
wrong_num = int((sentis != rest).sum())
all_num['total_num'] += total_num
all_num['wrong_num'] += wrong_num
senti_num[senti]['total_num'] += total_num
senti_num[senti]['wrong_num'] += wrong_num
random_id = random.randint(0, caps_tensor.size(0)-1)
caption = ' '.join([idx2word[idx] for idx in caps_tensor[random_id]])
pred_senti = rest_w[random_id]
att_weight = str(att_weights[random_id].detach().cpu().numpy().tolist())
test_case[senti].append([caption, pred_senti, att_weight])
tmp_total_num = 0
tmp_wrong_num = 0
for senti in senti_num:
tmp_total_num += senti_num[senti]['total_num']
tmp_wrong_num += senti_num[senti]['wrong_num']
assert tmp_total_num == all_num['total_num'] and tmp_wrong_num == all_num['wrong_num']
all_acc_rate = 100 - all_num['wrong_num'] / all_num['total_num'] * 100
senti_acc_rate = {}
for senti in senti_num:
senti_acc_rate[senti] = 100 - senti_num[senti]['wrong_num'] / senti_num[senti]['total_num'] * 100
json.dump(test_case, open(os.path.join(result_dir, 'test_case_%d_%.4f.json' % (epoch, all_acc_rate)), 'w'))
if previous_acc_rate is not None and all_acc_rate < previous_acc_rate:
lr = lr * 0.5
for param_group in optimizer.param_groups:
param_group['lr'] = lr
previous_acc_rate = all_acc_rate
print('train_loss: %.4f, all_acc_rate: %.4f, senti_acc_rate: %s' %
(train_loss, all_acc_rate, senti_acc_rate))
if epoch > -1:
chkpoint = {
'epoch': epoch,
'model': model.state_dict(),
'optimizer': optimizer.state_dict(),
'settings': opt.settings,
'idx2word': idx2word,
'sentiment_categories': opt.sentiment_categories,
'dataset_name': dataset_name,
'corpus_type': corpus_type,
}
checkpoint_path = os.path.join(checkpoint, 'model_%d_%.4f_%.4f_%s.pth' % (
epoch, train_loss, all_acc_rate, time.strftime('%m%d-%H%M')))
torch.save(chkpoint, checkpoint_path)
if __name__ == '__main__':
try:
opt = parse_opt()
train()
except BdbQuit:
sys.exit(1)
except Exception:
traceback.print_exc()
print('')
pdb.post_mortem()
sys.exit(1)