-
Notifications
You must be signed in to change notification settings - Fork 15
/
r_bert.py
executable file
·423 lines (365 loc) · 18.4 KB
/
r_bert.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
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
import argparse
import glob
import logging
import os
import sys
import random
import torch.nn as nn
import numpy as np
import torch
import socket
# wss
# import ptvsd
# Allow other computers to attach to ptvsd at this IP address and port.
# ptvsd.enable_attach(address=('192.168.11.2', 3000), redirect_output=True)
# Pause the program until a remote debugger is attached
# ptvsd.wait_for_attach()
from torch.utils.data import (DataLoader, RandomSampler, SequentialSampler,
TensorDataset)
from torch.utils.data.distributed import DistributedSampler
# from tensorboardX import SummaryWriter
from tqdm import tqdm, trange
from transformers import (WEIGHTS_NAME, BertConfig, BertTokenizer)
from transformers import AdamW, get_linear_schedule_with_warmup
from utils import (SEMEVAL_RELATION_LABELS, TACRED_RELATION_LABELS, compute_metrics,
convert_examples_to_features, output_modes, data_processors)
import torch.nn.functional as F
from argparse import ArgumentParser
from config import Config
from model import BertForSequenceClassification
logger = logging.getLogger(__name__)
additional_special_tokens = ["[E11]", "[E12]", "[E21]", "[E22]"]
def set_seed(seed):
random.seed(seed)
np.random.seed(seed)
torch.manual_seed(seed)
torch.cuda.manual_seed_all(seed)
def train(config, train_dataset, model, tokenizer):
""" Train the model """
config.train_batch_size = config.per_gpu_train_batch_size * \
max(1, config.n_gpu)
if config.local_rank == -1:
train_sampler = RandomSampler(train_dataset)
else:
DistributedSampler(train_dataset)
train_dataloader = DataLoader(
train_dataset, sampler=train_sampler, batch_size=config.train_batch_size)
if config.max_steps > 0:
t_total = config.max_steps
config.num_train_epochs = config.max_steps // (
len(train_dataloader) // config.gradient_accumulation_steps) + 1
else:
t_total = len(
train_dataloader) // config.gradient_accumulation_steps * config.num_train_epochs
# Prepare optimizer and schedule (linear warmup and decay)
no_decay = ['bias', 'LayerNorm.weight']
optimizer_grouped_parameters = [
{'params': [p for n, p in model.named_parameters()
if not any(nd in n for nd in no_decay)], 'weight_decay': config.weight_decay},
{'params': [p for n, p in model.named_parameters()
if any(nd in n for nd in no_decay)], 'weight_decay': 0.0}
]
optimizer = AdamW(optimizer_grouped_parameters,
lr=config.learning_rate, eps=config.adam_epsilon)
scheduler = get_linear_schedule_with_warmup(
optimizer, num_warmup_steps=config.warmup_steps, num_training_steps=t_total)
if config.n_gpu > 1:
model = torch.nn.DataParallel(model)
# Distributed training (should be after apex fp16 initialization)
if config.local_rank != -1:
model = torch.nn.parallel.DistributedDataParallel(model, device_ids=[config.local_rank],
output_device=config.local_rank,
find_unused_parameters=True)
# Train!
logger.info("***** Running training *****")
logger.info(" Num examples = %d", len(train_dataset))
logger.info(" Num Epochs = %d", config.num_train_epochs)
logger.info(" Instantaneous batch size per GPU = %d",
config.per_gpu_train_batch_size)
logger.info(" Total train batch size (w. parallel, distributed & accumulation) = %d",
config.train_batch_size * config.gradient_accumulation_steps
* (torch.distributed.get_world_size() if config.local_rank != -1 else 1))
logger.info(" Gradient Accumulation steps = %d",
config.gradient_accumulation_steps)
logger.info(" Total optimization steps = %d", t_total)
global_step = 0
tr_loss, logging_loss = 0.0, 0.0
model.zero_grad()
train_iterator = trange(int(config.num_train_epochs),
desc="Epoch", disable=config.local_rank not in [-1, 0])
# Added here for reproductibility (even between python 2 and 3)
set_seed(config.seed)
for _ in train_iterator:
epoch_iterator = tqdm(train_dataloader, desc="Iteration",
disable=config.local_rank not in [-1, 0])
for step, batch in enumerate(epoch_iterator):
model.train()
batch = tuple(t.to(config.device) for t in batch)
inputs = {'input_ids': batch[0],
'attention_mask': batch[1],
# XLM and RoBERTa don't use segment_ids
'token_type_ids': batch[2],
'labels': batch[3],
'e1_mask': batch[4],
'e2_mask': batch[5],
}
outputs = model(**inputs)
# model outputs are always tuple in pytorch-transformers (see doc)
loss = outputs[0]
if config.n_gpu > 1:
loss = loss.mean() # mean() to average on multi-gpu parallel training
if config.gradient_accumulation_steps > 1:
loss = loss / config.gradient_accumulation_steps
loss.backward()
torch.nn.utils.clip_grad_norm_(
model.parameters(), config.max_grad_norm)
tr_loss += loss.item()
if (step + 1) % config.gradient_accumulation_steps == 0:
optimizer.step()
scheduler.step() # Update learning rate schedule
model.zero_grad()
global_step += 1
if config.local_rank in [-1, 0] and config.logging_steps > 0 and global_step % config.logging_steps == 0:
# Log metrics
# Only evaluate when single GPU otherwise metrics may not average well
if config.local_rank == -1 and config.evaluate_during_training:
results = evaluate(config, model, tokenizer)
logging_loss = tr_loss
if config.local_rank in [-1, 0] and config.save_steps > 0 and global_step % config.save_steps == 0:
# Save model checkpoint
output_dir = os.path.join(
config.output_dir, 'checkpoint-{}'.format(global_step))
if not os.path.exists(output_dir):
os.makedirs(output_dir)
# Take care of distributed/parallel training
model_to_save = model.module if hasattr(
model, 'module') else model
model_to_save.save_pretrained(output_dir)
torch.save(config, os.path.join(
output_dir, 'training_config.bin'))
logger.info("Saving model checkpoint to %s", output_dir)
if config.max_steps > 0 and global_step > config.max_steps:
epoch_iterator.close()
break
if config.max_steps > 0 and global_step > config.max_steps:
train_iterator.close()
break
return global_step, tr_loss / global_step
def evaluate(config, model, tokenizer, prefix=""):
# Loop to handle MNLI double evaluation (matched, mis-matched)
eval_task = config.task_name
eval_output_dir = config.output_dir
results = {}
eval_dataset = load_and_cache_examples(
config, eval_task, tokenizer, evaluate=True)
if not os.path.exists(eval_output_dir) and config.local_rank in [-1, 0]:
os.makedirs(eval_output_dir)
config.eval_batch_size = config.per_gpu_eval_batch_size * \
max(1, config.n_gpu)
# Note that DistributedSampler samples randomly
eval_sampler = SequentialSampler(
eval_dataset) if config.local_rank == -1 else DistributedSampler(eval_dataset)
eval_dataloader = DataLoader(
eval_dataset, sampler=eval_sampler, batch_size=config.eval_batch_size, shuffle=False)
# Eval!
logger.info("***** Running evaluation {} *****".format(prefix))
logger.info(" Num examples = %d", len(eval_dataset))
logger.info(" Batch size = %d", config.eval_batch_size)
eval_loss = 0.0
nb_eval_steps = 0
preds = None
out_label_ids = None
for batch in tqdm(eval_dataloader, desc="Evaluating"):
model.eval()
batch = tuple(t.to(config.device) for t in batch)
with torch.no_grad():
inputs = {'input_ids': batch[0],
'attention_mask': batch[1],
# XLM and RoBERTa don't use segment_ids
'token_type_ids': batch[2],
'labels': batch[3],
'e1_mask': batch[4],
'e2_mask': batch[5],
}
outputs = model(**inputs)
tmp_eval_loss, logits = outputs[:2]
eval_loss += tmp_eval_loss.mean().item()
nb_eval_steps += 1
if preds is None:
preds = logits.detach().cpu().numpy()
out_label_ids = inputs['labels'].detach().cpu().numpy()
else:
preds = np.append(preds, logits.detach().cpu().numpy(), axis=0)
out_label_ids = np.append(
out_label_ids, inputs['labels'].detach().cpu().numpy(), axis=0)
eval_loss = eval_loss / nb_eval_steps
preds = np.argmax(preds, axis=1)
result = compute_metrics(eval_task, preds, out_label_ids)
results.update(result)
logger.info("***** Eval results {} *****".format(prefix))
for key in sorted(result.keys()):
logger.info(f"{key} = {result[key]}")
if config.task_name == "semeval":
output_eval_file = "eval/sem_res.txt"
with open(output_eval_file, "w") as writer:
for key in range(len(preds)):
writer.write("%d\t%s\n" %
(key+8001, str(SEMEVAL_RELATION_LABELS[preds[key]])))
elif config.task_name == "tacred":
output_eval_file = "eval/tac_res.txt"
with open(output_eval_file, "w") as writer:
for pred in preds:
writer.write(TACRED_RELATION_LABELS[pred])
writer.write("\n")
return result
def load_and_cache_examples(config, task, tokenizer, evaluate=False, test=False):
if config.local_rank not in [-1, 0] and not evaluate:
# Make sure only the first process in distributed training process the dataset, and the others will use the cache
torch.distributed.barrier()
processor = data_processors[config.task_name]()
output_mode = "classification"
# Load data features from cache or dataset file
evaluation_set_name = 'test' if test else 'dev'
cached_features_file = os.path.join(config.data_dir, 'cached_{}_{}_{}_{}'.format(
evaluation_set_name if evaluate else 'train',
list(filter(None, config.pretrained_model_name.split('/'))).pop(),
str(config.max_seq_len),
str(task)))
if os.path.exists(cached_features_file):
logger.info(f"Loading features from cached file {cached_features_file}")
features = torch.load(cached_features_file)
else:
logger.info("Creating features from dataset file at %s",
config.data_dir)
label_list = processor.get_labels()
examples = processor.get_dev_examples(
config.data_dir) if evaluate else processor.get_train_examples(config.data_dir)
features = convert_examples_to_features(
examples, label_list, config.max_seq_len, tokenizer, "classification", use_entity_indicator=config.use_entity_indicator)
if config.local_rank in [-1, 0]:
logger.info(f"Saving features into cached file {cached_features_file}")
torch.save(features, cached_features_file)
if config.local_rank == 0 and not evaluate:
# Make sure only the first process in distributed training process the dataset, and the others will use the cache
torch.distributed.barrier()
# Convert to Tensors and build dataset
all_input_ids = torch.tensor(
[f.input_ids for f in features], dtype=torch.long)
all_input_mask = torch.tensor(
[f.input_mask for f in features], dtype=torch.long)
all_segment_ids = torch.tensor(
[f.segment_ids for f in features], dtype=torch.long)
all_e1_mask = torch.tensor(
[f.e1_mask for f in features], dtype=torch.long) # add e1 mask
all_e2_mask = torch.tensor(
[f.e2_mask for f in features], dtype=torch.long) # add e2 mask
if output_mode == "classification":
all_label_ids = torch.tensor(
[f.label_id for f in features], dtype=torch.long)
elif output_mode == "regression":
all_label_ids = torch.tensor(
[f.label_id for f in features], dtype=torch.float)
dataset = TensorDataset(all_input_ids, all_input_mask,
all_segment_ids, all_label_ids, all_e1_mask, all_e2_mask)
return dataset
def main():
parser = ArgumentParser(
description="BERT for relation extraction (classification)")
parser.add_argument('--config', dest='config')
args = parser.parse_args()
config = Config(args.config)
if os.path.exists(config.output_dir) and os.listdir(config.output_dir) and config.train and not config.overwrite_output_dir:
raise ValueError(
"Output directory ({}) already exists and is not empty. Use --overwrite_output_dir to overcome.".format(config.output_dir))
# Setup CUDA, GPU & distributed training
if config.local_rank == -1 or config.no_cuda:
device = torch.device(
"cuda" if torch.cuda.is_available() and not config.no_cuda else "cpu")
# config.n_gpu = torch.cuda.device_count()
config.n_gpu = 1
else:
# Initializes the distributed backend which will take care of sychronizing nodes/GPUs
torch.cuda.set_device(config.local_rank)
device = torch.device("cuda", args.local_rank)
torch.distributed.init_process_group(backend='nccl')
config.n_gpu = 1
config.device = device
# Setup logging
logging.basicConfig(format='%(asctime)s - %(levelname)s - %(name)s - %(message)s',
datefmt='%m/%d/%Y %H:%M:%S',
level=logging.INFO if config.local_rank in [-1, 0] else logging.WARN)
logger.warning("Process rank: %s, device: %s, n_gpu: %s, distributed training: %s",
config.local_rank, device, config.n_gpu, bool(config.local_rank != -1))
# Set seed
set_seed(config.seed)
# Prepare task -- SemEval or TACRED
processor = data_processors[config.task_name]()
label_list = processor.get_labels()
num_labels = len(label_list)
# Load pretrained model and tokenizer
if config.local_rank not in [-1, 0]:
torch.distributed.barrier()
# Make sure only the first process in distributed training will download model & vocab
bertconfig = BertConfig.from_pretrained(
config.pretrained_model_name, num_labels=num_labels, finetuning_task=config.task_name)
do_lower_case = "-uncased" in config.pretrained_model_name
tokenizer = BertTokenizer.from_pretrained(
config.pretrained_model_name, do_lower_case=do_lower_case, additional_special_tokens=additional_special_tokens)
model = BertForSequenceClassification.from_pretrained(
config.pretrained_model_name, config=bertconfig)
if config.local_rank == 0:
# Make sure only the first process in distributed training will download model & vocab
torch.distributed.barrier()
model.to(config.device)
# logger.info("Training/evaluation parameters %s", config)
# Training
if config.train:
train_dataset = load_and_cache_examples(
config, config.task_name, tokenizer, evaluate=False)
global_step, tr_loss = train(config, train_dataset, model, tokenizer)
logger.info(" global_step = %s, average loss = %s",
global_step, tr_loss)
# Saving best-practices: if you use defaults names for the model, you can reload it using from_pretrained()
if config.train and (config.local_rank == -1 or torch.distributed.get_rank() == 0):
# Create output directory if needed
if not os.path.exists(config.output_dir) and config.local_rank in [-1, 0]:
os.makedirs(config.output_dir)
logger.info("Saving model checkpoint to %s", config.output_dir)
# Save a trained model, configuration and tokenizer using `save_pretrained()`.
# They can then be reloaded using `from_pretrained()`
# Take care of distributed/parallel training
model_to_save = model.module if hasattr(model, 'module') else model
model_to_save.save_pretrained(config.output_dir)
tokenizer.save_pretrained(config.output_dir)
# Good practice: save your training arguments together with the trained model
torch.save(config, os.path.join(
config.output_dir, 'training_config.bin'))
# Load a trained model and vocabulary that you have fine-tuned
model = BertForSequenceClassification.from_pretrained(config.output_dir)
tokenizer = BertTokenizer.from_pretrained(
config.output_dir, do_lower_case=do_lower_case, additional_special_tokens=additional_special_tokens)
model.to(config.device)
# Evaluation
results = {}
if config.eval and config.local_rank in [-1, 0]:
tokenizer = BertTokenizer.from_pretrained(
config.output_dir, do_lower_case=do_lower_case, additional_special_tokens=additional_special_tokens)
checkpoints = [config.output_dir]
if config.eval_all_checkpoints:
checkpoints = list(os.path.dirname(c) for c in sorted(
glob.glob(config.output_dir + '/**/' + WEIGHTS_NAME, recursive=True)))
logging.getLogger("pytorch_transformers.modeling_utils").setLevel(
logging.WARN) # Reduce logging
logger.info("Evaluate the following checkpoints: %s", checkpoints)
for checkpoint in checkpoints:
global_step = checkpoint.split(
'-')[-1] if len(checkpoints) > 1 else ""
model = BertForSequenceClassification.from_pretrained(checkpoint)
model.to(config.device)
result = evaluate(config, model, tokenizer, prefix=global_step)
result = dict((k + '_{}'.format(global_step), v)
for k, v in result.items())
results.update(result)
return results
if __name__ == "__main__":
main()