-
Notifications
You must be signed in to change notification settings - Fork 0
/
main.py
306 lines (256 loc) · 11.9 KB
/
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
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from functools import partial
import argparse
import os
import random
import time
import numpy as np
import paddle
import paddle.nn.functional as F
import paddlenlp as ppnlp
from paddlenlp.data import Stack, Tuple, Pad
from paddlenlp.datasets import load_dataset
from paddlenlp.transformers import LinearDecayWithWarmup
# yapf: disable
parser = argparse.ArgumentParser()
parser.add_argument("--train_path", default="./dataset/train_hyper.tsv", type=str, help="The train dataset path.")
parser.add_argument("--dev_path", default="./dataset/dev_hyper.tsv", type=str, help="The dev dataset path.")
parser.add_argument("--test_path", default="./dataset/test_hyper.tsv", type=str, help="The test dataset path.")
parser.add_argument("--save_dir", default="./checkpoint", type=str,
help="The output directory where the model checkpoints will be written.")
parser.add_argument("--max_seq_length", default=20, type=int,
help="The maximum total input sequence length after tokenization. "
"Sequences longer than this will be truncated, sequences shorter will be padded.")
parser.add_argument("--batch_size", default=256, type=int, help="Batch size per GPU/CPU for training.")
parser.add_argument("--learning_rate", default=5e-5, type=float, help="The initial learning rate for Adam.")
parser.add_argument("--weight_decay", default=0.0, type=float, help="Weight decay if we apply some.")
parser.add_argument("--epochs", default=3, type=int, help="Total number of training epochs to perform.")
parser.add_argument("--warmup_proportion", default=0.0, type=float,
help="Linear warmup proption over the training process.")
parser.add_argument("--init_from_ckpt", type=str, default=None, help="The path of checkpoint to be loaded.")
parser.add_argument("--seed", type=int, default=1000, help="random seed for initialization")
parser.add_argument('--device', choices=['cpu', 'gpu', 'xpu'], default="gpu",
help="Select which device to train model, defaults to gpu.")
args = parser.parse_args()
def set_seed(seed):
"""sets random seed"""
random.seed(seed)
np.random.seed(seed)
paddle.seed(seed)
@paddle.no_grad()
def evaluate(model, criterion, metric, data_loader):
"""
Given a dataset, it evals model and computes the metric.
Args:
model(obj:`paddle.nn.Layer`): A model to classify texts.
data_loader(obj:`paddle.io.DataLoader`): The dataset loader which generates batches.
criterion(obj:`paddle.nn.Layer`): It can compute the loss.
metric(obj:`paddle.metric.Metric`): The evaluation metric.
"""
model.eval()
metric.reset()
losses = []
for batch in data_loader:
print('pass')
input_ids, token_type_ids, labels = batch
logits = model(input_ids, token_type_ids)
loss = criterion(logits, labels)
losses.append(loss.numpy())
correct = metric.compute(logits, labels)
metric.update(correct)
accu = metric.accumulate()
print("eval loss: %.5f, accu: %.5f" % (np.mean(losses), accu))
model.train()
metric.reset()
def convert_example(example, tokenizer, max_seq_length=512, is_test=False):
"""
Builds model inputs from a sequence or a pair of sequence for sequence classification tasks
by concatenating and adding special tokens. And creates a mask from the two sequences passed
to be used in a sequence-pair classification task.
A BERT sequence has the following format:
- single sequence: ``[CLS] X [SEP]``
- pair of sequences: ``[CLS] A [SEP] B [SEP]``
A BERT sequence pair mask has the following format:
::
0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1
| first sequence | second sequence |
If only one sequence, only returns the first portion of the mask (0's).
Args:
example(obj:`list[str]`): List of input data, containing text and label if it have label.
tokenizer(obj:`PretrainedTokenizer`): This tokenizer inherits from :class:`~paddlenlp.transformers.PretrainedTokenizer`
which contains most of the methods. Users should refer to the superclass for more information regarding methods.
max_seq_len(obj:`int`): The maximum total input sequence length after tokenization.
Sequences longer than this will be truncated, sequences shorter will be padded.
is_test(obj:`False`, defaults to `False`): Whether the example contains label or not.
Returns:
input_ids(obj:`list[int]`): The list of token ids.
token_type_ids(obj: `list[int]`): List of sequence pair mask.
label(obj:`numpy.array`, data type of int64, optional): The input label if not is_test.
"""
encoded_inputs = tokenizer(text=example["hypo"], text_pair=example["hyper"], max_seq_len=max_seq_length)
input_ids = encoded_inputs["input_ids"]
token_type_ids = encoded_inputs["token_type_ids"]
if not is_test:
label = np.array([example["label"]], dtype="int64")
return input_ids, token_type_ids, label
else:
return input_ids, token_type_ids
def read(file_path): # 0 表示是上下位关系,1表示不是上下位关系
with open(file_path, 'r', encoding='utf-8') as f_in:
for cnt, line in enumerate(f_in):
line = line.strip()
if not line:
continue
res_lst = line.split('\t')
if len(res_lst) != 3:
print('数据集错误,行数%d' % cnt)
return None
hypo, hyper, label = res_lst
hypo, hyper = hypo.replace('_', ''), hyper.replace('_', '') # 去掉数据集中的下划线
label = 0 if label == 'hyper' else 1
yield {'hypo': hypo, 'hyper': hyper, 'label': label}
def create_dataloader(dataset,
mode='train',
batch_size=1,
batchify_fn=None,
trans_fn=None):
if trans_fn:
dataset = dataset.map(trans_fn)
shuffle = True if mode == 'train' else False
if mode == 'train':
batch_sampler = paddle.io.DistributedBatchSampler(
dataset, batch_size=batch_size, shuffle=shuffle)
else:
batch_sampler = paddle.io.BatchSampler(
dataset, batch_size=batch_size, shuffle=shuffle)
return paddle.io.DataLoader(
dataset=dataset,
batch_sampler=batch_sampler,
collate_fn=batchify_fn,
return_list=True)
def do_train():
paddle.set_device(args.device)
rank = paddle.distributed.get_rank()
if paddle.distributed.get_world_size() > 1:
paddle.distributed.init_parallel_env()
set_seed(args.seed)
train_ds = load_dataset(read, file_path=args.train_path, lazy=False)
dev_ds = load_dataset(read, file_path=args.dev_path, lazy=False)
model = ppnlp.transformers.BertForSequenceClassification.from_pretrained('bert-base-chinese', num_class=2)
tokenizer = ppnlp.transformers.BertTokenizer.from_pretrained('bert-base-chinese')
trans_func = partial(
convert_example,
tokenizer=tokenizer,
max_seq_length=args.max_seq_length)
batchify_fn = lambda samples, fn=Tuple(
Pad(axis=0, pad_val=tokenizer.pad_token_id), # input
Pad(axis=0, pad_val=tokenizer.pad_token_type_id), # segment
Stack(dtype="int64") # label
): [data for data in fn(samples)]
train_data_loader = create_dataloader(
train_ds,
mode='train',
batch_size=args.batch_size,
batchify_fn=batchify_fn,
trans_fn=trans_func)
dev_data_loader = create_dataloader(
dev_ds,
mode='dev',
batch_size=args.batch_size,
batchify_fn=batchify_fn,
trans_fn=trans_func)
if args.init_from_ckpt and os.path.isfile(args.init_from_ckpt):
state_dict = paddle.load(args.init_from_ckpt)
model.set_dict(state_dict)
model = paddle.DataParallel(model)
num_training_steps = len(train_data_loader) * args.epochs
lr_scheduler = LinearDecayWithWarmup(args.learning_rate, num_training_steps,
args.warmup_proportion)
# Generate parameter names needed to perform weight decay.
# All bias and LayerNorm parameters are excluded.
decay_params = [
p.name for n, p in model.named_parameters()
if not any(nd in n for nd in ["bias", "norm"])
]
optimizer = paddle.optimizer.AdamW(
learning_rate=lr_scheduler,
parameters=model.parameters(),
weight_decay=args.weight_decay,
apply_decay_param_fun=lambda x: x in decay_params)
criterion = paddle.nn.loss.CrossEntropyLoss()
metric = paddle.metric.Accuracy()
global_step = 0
tic_train = time.time()
for epoch in range(1, args.epochs + 1):
for step, batch in enumerate(train_data_loader, start=1):
input_ids, token_type_ids, labels = batch
logits = model(input_ids, token_type_ids)
loss = criterion(logits, labels)
probs = F.softmax(logits, axis=1)
correct = metric.compute(probs, labels)
metric.update(correct)
acc = metric.accumulate()
global_step += 1
if global_step % 10 == 0 and rank == 0:
print(
"global step %d, epoch: %d, batch: %d, loss: %.5f, accu: %.5f, speed: %.2f step/s"
% (global_step, epoch, step, loss, acc,
10 / (time.time() - tic_train)))
tic_train = time.time()
loss.backward()
optimizer.step()
lr_scheduler.step()
optimizer.clear_grad()
if global_step % 100 == 0 and rank == 0:
save_dir = os.path.join(args.save_dir, "model_%d" % global_step)
if not os.path.exists(save_dir):
os.makedirs(save_dir)
evaluate(model, criterion, metric, dev_data_loader)
model._layers.save_pretrained(save_dir)
tokenizer.save_pretrained(save_dir)
def do_eval():
paddle.set_device(args.device)
if paddle.distributed.get_world_size() > 1:
paddle.distributed.init_parallel_env()
set_seed(args.seed)
test_ds = load_dataset(read, file_path=args.test_path, lazy=False)
model = ppnlp.transformers.BertForSequenceClassification.from_pretrained('bert-base-chinese', num_class=2)
tokenizer = ppnlp.transformers.BertTokenizer.from_pretrained('bert-base-chinese')
trans_func = partial(
convert_example,
tokenizer=tokenizer,
max_seq_length=args.max_seq_length)
batchify_fn = lambda samples, fn=Tuple(
Pad(axis=0, pad_val=tokenizer.pad_token_id), # input
Pad(axis=0, pad_val=tokenizer.pad_token_type_id), # segment
Stack(dtype="int64") # label
): [data for data in fn(samples)]
test_data_loader = create_dataloader(
test_ds,
mode='test',
batch_size=args.batch_size,
batchify_fn=batchify_fn,
trans_fn=trans_func
)
if args.init_from_ckpt and os.path.isfile(args.init_from_ckpt):
state_dict = paddle.load(args.init_from_ckpt)
model.set_dict(state_dict)
model = paddle.DataParallel(model)
criterion = paddle.nn.loss.CrossEntropyLoss()
metric = paddle.metric.Accuracy()
evaluate(model, criterion, metric, test_data_loader)
if __name__ == "__main__":
do_train()
# do_eval()