-
Notifications
You must be signed in to change notification settings - Fork 0
/
joint_main.py
196 lines (178 loc) · 7.65 KB
/
joint_main.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
# @Author : guopeiming
# @Contact : guopeiming.gpm@{qq, gmail}.com
import os
import time
import torch
import random
import argparse
import datetime
import numpy as np
from utils.optim import Optim
from utils import joint_evaluate
from model.JointModel import JointModel
from torch.utils.data import DataLoader
from config.joint_args import parse_args
from utils.trees import InternalTreebankNode
from typing import Tuple, List, Set, Union, Dict
from utils.joint_dataset import load_data, batch_filter, batch_spliter, write_joint_data
def preprocess() -> argparse.Namespace:
"""
preprocess of training
:return: config args
"""
print('preprocessing starts...\n')
# ====== parse arguments ====== #
args = parse_args()
# ====== set random seed ====== #
random.seed(args.seed)
np.random.seed(args.seed)
torch.manual_seed(args.seed)
torch.cuda.manual_seed_all(args.seed)
# ====== save path ====== #
now_time = datetime.datetime.now().strftime('%Y-%m-%d-%H-%M-%S')
args.save_path = os.path.join('./logs/', 'my_log-' + now_time)
if not os.path.exists(args.save_path) and not args.debug:
os.makedirs(args.save_path)
# ====== cuda enable ====== #
os.environ['CUDA_VISIBLE_DEVICES'] = str(args.gpuid)
args.device = torch.device('cuda') if args.cuda and torch.cuda.is_available() else torch.device('cpu')
# ====== others ====== #
os.environ['TOKENIZERS_PARALLELISM'] = 'false'
torch.set_num_threads(6)
print(args, end='\n\n')
return args
def postprocess(args: argparse.Namespace, start: float):
exe_time = time.time() - start
print('Executing time: %dh:%dm:%ds.' % (exe_time//3600, (exe_time//60) % 60, exe_time % 60))
print('training ends.')
@torch.no_grad()
def eval_model(model: torch.nn.Module, dataset: DataLoader, language: str, subword: str, DATASET_MAX_SNT_LENGTH: int,
BATCH_MAX_SNT_LENGTH: int, evalb_path: str, type_: str)\
-> Tuple[joint_evaluate.JointFScore,
Dict[str, List[Union[InternalTreebankNode, Set[Tuple[str, Tuple[int, int]]]]]]]:
trees_pred, trees_gold = [], []
for insts in dataset:
model.eval()
insts, _, max_len = batch_filter(insts, language, DATASET_MAX_SNT_LENGTH)
insts_list = batch_spliter(insts, max_len, BATCH_MAX_SNT_LENGTH)
for insts in insts_list:
trees_batch_pred, _ = model(insts)
trees_batch_gold = insts['joint_gold_trees']
trees_pred.extend(trees_batch_pred)
trees_gold.extend(trees_batch_gold)
assert len(trees_pred) == len(trees_gold)
joint_fscore, res_dict = joint_evaluate.cal_performance(language, subword, evalb_path, trees_gold, trees_pred)
print('Model performance in %s dataset: JointFScore: %s' % (type_, joint_fscore))
torch.cuda.empty_cache()
return joint_fscore, res_dict
def main():
# ====== preprocess ====== #
args = preprocess()
# ====== Loading dataset ====== #
train_data, dev_data, test_data, joint_vocabs, parsing_vocabs = load_data(
args.joint_input, args.parsing_input, args.batch_size, args.accum_steps, args.shuffle, args.num_workers,
args.drop_last
)
# cross_labels_idx = generate_cross_labels_idx(vocabs['labels'])
# ======= Preparing Model ======= #
print("\nModel Preparing starts...")
model = JointModel(
joint_vocabs,
parsing_vocabs,
# cross_labels_idx,
# Embedding
args.subword,
args.bert_path,
args.transliterate,
args.d_model,
args.partition,
args.position_emb_dropout,
args.bert_emb_dropout,
args.emb_dropout,
# Encoder
args.layer_num,
args.hidden_dropout,
args.attention_dropout,
args.dim_ff,
args.nhead,
args.kqv_dim,
# classifier
args.label_hidden,
# loss
args.lambda_scaler,
args.alpha_scaler,
args.language,
args.device
)
if args.pretrain_model_path is not False:
model.load_pretrain_model(args.pretrain_model_path)
if args.cuda:
model = model.cuda()
# print(model, end='\n\n\n')
optimizer = Optim(model, args.optim, args.lr, args.lr_fine_tune, args.warmup_steps, args.lr_decay_factor,
args.weight_decay, args.clip_grad, args.clip_grad_max_norm)
optimizer.zero_grad()
# if args.freeze_bert:
# optimizer.set_freeze_by_idxs([str(num) for num in range(0, config.freeze_bert_layers)], True)
# optimizer.free_embeddings()
# optimizer.freeze_pooler()
# print('freeze model of BERT %d layers' % config.freeze_bert_layers)
# ========= Training ========= #
print('Training starts...')
start = time.time()
steps, loss_value, total_batch_size = 1, 0., 0
best_dev, best_test = None, None
patience = args.patience
for epoch_i in range(1, args.epoch):
for batch_i, insts in enumerate(train_data, start=1):
model.train()
insts, batch_size, max_len = batch_filter(insts, args.language, args.DATASET_MAX_SNT_LENGTH)
insts_list = batch_spliter(insts, max_len, args.BATCH_MAX_SNT_LENGTH)
total_batch_size += batch_size
for insts in insts_list:
loss = model(insts)
if loss.item() > 0.:
loss.backward()
loss_value += loss.item()
assert not isinstance(loss_value, torch.Tensor), 'GPU memory leak'
if steps % args.accum_steps == 0:
optimizer.step()
optimizer.zero_grad()
if steps % (args.accum_steps * args.log_interval) == 0:
print(
'[%d/%d], [%d/%d] Loss: %.05f' %
(epoch_i, args.epoch, batch_i//args.accum_steps, len(train_data)//args.accum_steps,
loss_value/total_batch_size), flush=True
)
loss_value, total_batch_size = 0., 0
torch.cuda.empty_cache()
if steps % (args.accum_steps * args.eval_interval) == 0:
print('model evaluating starts...', flush=True)
joint_fscore_dev, res_data_dev = eval_model(
model, dev_data, args.language, args.subword, args.DATASET_MAX_SNT_LENGTH,
args.BATCH_MAX_SNT_LENGTH, args.evalb_path, 'dev'
)
joint_fscore_test, res_data_test = eval_model(
model, test_data, args.language, args.subword, args.DATASET_MAX_SNT_LENGTH,
args.BATCH_MAX_SNT_LENGTH, args.evalb_path, 'test'
)
if best_dev is None or joint_fscore_dev.parsing_f > best_dev.parsing_f:
best_dev, best_test = joint_fscore_dev, joint_fscore_test
patience = args.patience
write_joint_data(args.save_path, res_data_dev, 'dev')
write_joint_data(args.save_path, res_data_test, 'test')
print('best performance:\ndev: %s\ntest: %s' % (best_dev, best_test))
print('model evaluating ends...', flush=True)
del res_data_dev, res_data_test
if args.debug:
exit(0)
steps += 1
if args.early_stop:
patience -= 1
if patience < 0:
print('early stop')
break
# ====== postprocess ====== #
postprocess(args, start)
if __name__ == '__main__':
main()