-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathmain_transformer.py
1110 lines (1004 loc) · 51.1 KB
/
main_transformer.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
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
import glob
import json
import logging
import os
import random
import sys
import numpy as np
import torch
from sklearn.model_selection import train_test_split
from torch.utils.data import DataLoader, SequentialSampler, TensorDataset, ConcatDataset
from tqdm import tqdm
from transformers import (
WEIGHTS_NAME,
AlbertConfig,
AlbertForSequenceClassification,
AlbertTokenizer,
# BertConfig,
# BertForSequenceClassification,
BertTokenizer,
# DistilBertConfig,
# DistilBertForSequenceClassification,
DistilBertTokenizer,
FlaubertConfig,
FlaubertForSequenceClassification,
FlaubertTokenizer,
RobertaConfig,
# RobertaForSequenceClassification,
RobertaTokenizer,
XLMConfig,
XLMForSequenceClassification,
XLMRobertaConfig,
XLMRobertaForSequenceClassification,
XLMRobertaTokenizer,
XLMTokenizer,
XLNetConfig,
# XLNetForSequenceClassification,
XLNetTokenizer,
)
from transformers import (
AutoConfig,
AutoModelForSequenceClassification,
AutoTokenizer,
EvalPrediction,
HfArgumentParser,
PretrainedConfig,
Trainer,
TrainingArguments,
default_data_collator,
set_seed,
)
from src.transformers.configuration_distilbert import DistilBertConfig
from src.transformers.modeling_distilbert import DistilBertForSequenceClassification
sys.path.append("../../")
sys.path.append(os.path.dirname(os.path.dirname(os.path.dirname(__file__))))
from src.general import acc, f1_macro, precision_macro, recall_macro, create_dir
from src.glue.run_glue import compute_metrics, train
from src.metrics import uncertainty_metrics
from src.temperature_scaling import tune_temperature
from src.transformers.configuration_bert import BertConfig
from src.transformers.modeling_bert import BertForSequenceClassification
from src.transformers.processors import processors, output_modes, convert_examples_to_features
from sys_config import CACHE_DIR, DATA_DIR, CKPT_DIR, RES_DIR
try:
from torch.utils.tensorboard import SummaryWriter
except ImportError:
from tensorboardX import SummaryWriter
logger = logging.getLogger(__name__)
task_to_keys = {
"cola": ("sentence", None),
"mnli": ("premise", "hypothesis"),
"mrpc": ("sentence1", "sentence2"),
"qnli": ("question", "sentence"),
"qqp": ("question1", "question2"),
"rte": ("sentence1", "sentence2"),
"sst2": ("sentence", None),
"stsb": ("sentence1", "sentence2"),
"wnli": ("sentence1", "sentence2"),
}
# ALL_MODELS = sum(
# (
# tuple(conf.pretrained_config_archive_map.keys())
# for conf in (
# BertConfig,
# XLNetConfig,
# XLMConfig,
# RobertaConfig,
# DistilBertConfig,
# AlbertConfig,
# XLMRobertaConfig,
# FlaubertConfig,
# )
# ),
# (),
# )
# todo add more models
MODEL_CLASSES = {
"bert": (BertConfig, BertForSequenceClassification, BertTokenizer),
# "xlnet": (XLNetConfig, XLNetForSequenceClassification, XLNetTokenizer),
# "xlm": (XLMConfig, XLMForSequenceClassification, XLMTokenizer),
# "roberta": (RobertaConfig, RobertaForSequenceClassification, RobertaTokenizer),
"distilbert": (DistilBertConfig, DistilBertForSequenceClassification, DistilBertTokenizer),
# "albert": (AlbertConfig, AlbertForSequenceClassification, AlbertTokenizer),
# "xlmroberta": (XLMRobertaConfig, XLMRobertaForSequenceClassification, XLMRobertaTokenizer),
# "flaubert": (FlaubertConfig, FlaubertForSequenceClassification, FlaubertTokenizer),
}
def get_glue_dataset(args, data_dir, task, model_type, evaluate=False, test=False, contrast=False, ood=False):
"""
Loads original glue dataset.
:param data_dir: path ../data/[task]
:param task: glue task ("cola", "mnli", "mnli-mm", "mrpc", "sst-2", "sts-b", "qqp", "qnli", "rte", "wnli",
"ag_news", "dbpedia", "trec-6")
:param model_type: the type of the model we use (e.g. 'bert')
:param train: if True return dev set
:param evaluate: if True return dev set
:param test: if True return test set
:param test: if True return contrast set
:return:
"""
create_dir(data_dir)
processor = processors[task]()
output_mode = output_modes[task]
tokenizer = AutoTokenizer.from_pretrained(
args.tokenizer_name if args.tokenizer_name else args.model_name_or_path,
cache_dir=args.cache_dir,
use_fast=args.use_fast_tokenizer,
)
# Dataset
# Load data features from cache or dataset file
if test:
filename = "cached_{}_{}_original".format("test_contrast", str(task)) if contrast else "cached_{}_{}_original".format("test", str(task))
if ood: filename += '_ood'
cached_dataset = os.path.join(
data_dir,
filename
)
else:
if evaluate and contrast:
filename = "cached_{}_{}_original".format("dev_contrast", str(task))
else:
filename = "cached_{}_{}_original".format("dev" if evaluate else "train", str(task))
cached_dataset = os.path.join(
data_dir,
filename,
)
if os.path.exists(cached_dataset):
logger.info("Loading dataset from cached file %s", cached_dataset)
dataset = torch.load(cached_dataset)
else:
logger.info("Creating dataset from dataset file at %s", data_dir)
label_list = processor.get_labels()
if task in ["mnli", "mnli-mm"] and model_type in ["roberta", "xlmroberta"]:
# HACK(label indices are swapped in RoBERTa pretrained model)
label_list[1], label_list[2] = label_list[2], label_list[1]
if test:
if ood:
examples = (
processor.get_test_examples_ood(data_dir)
)
else:
examples = (
processor.get_contrast_examples("test") if contrast else processor.get_test_examples(data_dir)
)
else:
if evaluate and contrast:
examples = (
processor.get_contrast_examples("dev")
)
else:
examples = (
processor.get_dev_examples(data_dir) if evaluate else processor.get_train_examples(data_dir)
)
features = convert_examples_to_features(
examples,
tokenizer,
label_list=label_list,
max_length=args.max_seq_length,
output_mode=output_mode,
pad_on_left=bool(args.model_type in ["xlnet"]), # pad on the left for xlnet
pad_token=tokenizer.convert_tokens_to_ids([tokenizer.pad_token])[0],
pad_token_segment_id=4 if args.model_type in ["xlnet"] else 0,
)
if task in ['sst-2', 'cola', 'ag_news', 'dbpedia', 'trec-6', 'imdb']:
X = [x.text_a for x in examples]
elif task in ['mrpc', 'mnli', 'qnli', 'rte', 'qqp', 'wnli']:
X = list(zip([x.text_a for x in examples], [x.text_b for x in examples]))
# elif task == 'mnli':
# X = list(zip([x.text_a for x in examples], [x.text_b for x in examples]))
else:
print(task)
NotImplementedError
y = [x.label for x in examples]
dataset = [X, y]
logger.info("Saving dataset into cached file %s", cached_dataset)
torch.save(dataset, cached_dataset)
# Save Tensor Dataset
if test:
filename = "test_contrast" if contrast else "test"
if ood: filename += '_ood'
features_dataset = os.path.join(
args.data_dir,
"cached_{}_{}_{}_{}_original".format(
# "test",
filename,
list(filter(None, args.model_name_or_path.split("/"))).pop(),
str(args.max_seq_length),
str(task),
),
)
else:
filename = "dev" if evaluate else "train"
if contrast: filename += "_contrast"
features_dataset = os.path.join(
args.data_dir,
"cached_{}_{}_{}_{}_original".format(
# "dev" if evaluate else "train",
filename,
list(filter(None, args.model_name_or_path.split("/"))).pop(),
str(args.max_seq_length),
str(task),
),
)
torch.save(features, features_dataset)
return dataset
def get_glue_tensor_dataset(X_inds, args, task, tokenizer, train=False,
evaluate=False, test=False, augm=False, X_orig=None, X_augm=None, y_augm=None,
augm_features=None, dpool=False,
contrast=False, contrast_ori=False,
ood=False, data_dir=None):
"""
Load tensor dataset (not original/raw).
:param X_inds: list of indices to keep in the dataset (if None keep all)
:param args: args
:param task: task name
:param tokenizer: tokenizer
:param train: if True train dataset
:param evaluate: if True dev dataset
:param test: if True test dataset
:param augm: if True augmented dataset
:param X: augmented text (inputs)
:param y: augmented labels (original if augmentation of labeled data) if unlabeled ?
:return:
"""
if args.local_rank not in [-1, 0] and not evaluate:
torch.distributed.barrier() # Make sure only the first process in distributed training process the dataset, and the others will use the cache
if data_dir is None: data_dir = args.data_dir
processor = processors[task.lower()]()
output_mode = output_modes[task.lower()]
# Load data features from cache or dataset file
if test:
prefix = "test"
elif evaluate:
prefix="dev"
elif train:
prefix="train"
elif augm:
prefix="augm"
else:
prefix="???"
if contrast: prefix += "_contrast"
if contrast_ori: prefix += "_contrast_ori"
if ood: prefix += "_ood"
cached_features_file = os.path.join(
data_dir,
"cached_{}_{}_{}_{}_original".format(
prefix,
list(filter(None, args.model_name_or_path.split("/"))).pop(),
str(args.max_seq_length),
str(task),
),
)
if os.path.exists(cached_features_file): #and not args.overwrite_cache:
logger.info("Loading features from cached file %s", cached_features_file)
features = torch.load(cached_features_file)
if X_inds is not None:
logger.info("Selecting subsample...")
features = list(np.array(features)[X_inds])
else:
logger.info("Creating features from dataset file at %s", args.data_dir)
label_list = processor.get_labels()
if task in ["mnli", "mnli-mm"] and args.model_type in ["roberta", "xlmroberta"]:
# HACK(label indices are swapped in RoBERTa pretrained model)
label_list[1], label_list[2] = label_list[2], label_list[1]
if test:
if ood:
examples = (processor.get_test_examples(data_dir, ood))
else:
examples = (processor.get_contrast_examples("test", contrast_ori) if (contrast or contrast_ori) else processor.get_test_examples(data_dir))
elif evaluate:
examples = (processor.get_contrast_examples("dev") if contrast else processor.get_dev_examples(args.data_dir))
elif train:
examples = (processor.get_train_examples(args.data_dir))
elif augm:
if dpool:
augm_examples = (processor.get_augm_examples(X_augm, y_augm))
features = convert_examples_to_features(
augm_examples,
tokenizer,
label_list=label_list,
max_length=args.max_seq_length,
output_mode=output_mode,
pad_on_left=bool(args.model_type in ["xlnet"]), # pad on the left for xlnet
pad_token=tokenizer.convert_tokens_to_ids([tokenizer.pad_token])[0],
pad_token_segment_id=4 if args.model_type in ["xlnet"] else 0,
)
# Convert to Tensors and build dataset
all_input_ids = torch.tensor([f.input_ids for f in features], dtype=torch.long)
all_attention_mask = torch.tensor([f.attention_mask for f in features], dtype=torch.long)
all_token_type_ids = torch.tensor([f.token_type_ids for f in features], dtype=torch.long)
if output_mode == "classification":
all_labels = torch.tensor([f.label for f in features], dtype=torch.long)
elif output_mode == "regression":
all_labels = torch.tensor([f.label for f in features], dtype=torch.float)
dataset = TensorDataset(all_input_ids, all_attention_mask, all_token_type_ids, all_labels)
return dataset
else:
if X_orig is None: # DA for supervised learning
examples = (processor.get_augm_examples(X_augm, y_augm))
else: # DA for semi-supervised learning (consistency loss)
orig_examples = (processor.get_train_examples(args.data_dir))
orig_examples = [np.array(orig_examples)[i] for i in X_inds]
augm_examples = (processor.get_augm_examples(X_augm, y_augm))
assert len(orig_examples)==len(augm_examples), "orig len {}, augm len {}".format(len(orig_examples),
len(augm_examples))
all_input_ids = []
all_attention_mask = []
all_token_type_ids = []
all_labels = []
for examples in [orig_examples, augm_examples]:
features = convert_examples_to_features(
examples,
tokenizer,
label_list=label_list,
max_length=args.max_seq_length,
output_mode=output_mode,
pad_on_left=bool(args.model_type in ["xlnet"]), # pad on the left for xlnet
pad_token=tokenizer.convert_tokens_to_ids([tokenizer.pad_token])[0],
pad_token_segment_id=4 if args.model_type in ["xlnet"] else 0,
)
all_input_ids.append(torch.tensor([f.input_ids for f in features], dtype=torch.long))
all_attention_mask.append(torch.tensor([f.attention_mask for f in features], dtype=torch.long))
all_token_type_ids.append(torch.tensor([f.token_type_ids for f in features], dtype=torch.long))
if output_mode == "classification":
all_labels.append(torch.tensor([f.label for f in features], dtype=torch.long))
elif output_mode == "regression":
all_labels.append(torch.tensor([f.label for f in features], dtype=torch.float))
dataset = TensorDataset(all_input_ids[0], all_attention_mask[0], all_token_type_ids[0],
all_input_ids[1], all_attention_mask[1], all_token_type_ids[1],
all_labels[0])
return dataset
################################################################
if X_inds is not None:
examples = list(np.array(examples)[X_inds])
################################################################
features = convert_examples_to_features(
examples,
tokenizer,
label_list=label_list,
max_length=args.max_seq_length,
output_mode=output_mode,
pad_on_left=bool(args.model_type in ["xlnet"]), # pad on the left for xlnet
pad_token=tokenizer.convert_tokens_to_ids([tokenizer.pad_token])[0],
pad_token_segment_id=4 if args.model_type in ["xlnet"] else 0,
)
# if args.local_rank in [-1, 0]:
# logger.info("Saving features into cached file %s", cached_features_file)
# torch.save(features, cached_features_file)
if augm_features is not None:
# append train + augmented features (for DA supervised learning)
features = features + augm_features
if augm and augm_features is None:
# return augmented features (to later append with trainset for DA supervised learning)
return features
else:
if args.local_rank == 0 and not evaluate:
torch.distributed.barrier() # Make sure only the first process in distributed training process the dataset, and the others will use the cache
# Convert to Tensors and build dataset
all_input_ids = torch.tensor([f.input_ids for f in features], dtype=torch.long)
all_attention_mask = torch.tensor([f.attention_mask for f in features], dtype=torch.long)
if features[0].token_type_ids is not None:
all_token_type_ids = torch.tensor([f.token_type_ids for f in features], dtype=torch.long)
if output_mode == "classification":
all_labels = torch.tensor([f.label for f in features], dtype=torch.long)
elif output_mode == "regression":
all_labels = torch.tensor([f.label for f in features], dtype=torch.float)
if features[0].token_type_ids is not None:
dataset = TensorDataset(all_input_ids, all_attention_mask, all_token_type_ids, all_labels)
else:
dataset = TensorDataset(all_input_ids, all_attention_mask, all_labels)
if X_inds is not None and augm_features is None:
assert len(dataset) == len(X_inds), 'dataset {}, X_inds {}'.format(len(dataset), len(X_inds))
return dataset
def my_evaluate(eval_dataset, args, model, prefix="", mc_samples=None,
return_bert_embs=False):
"""
Evaluate model using 'eval_dataset'.
:param eval_dataset: tensor dataset
:param args:
:param model:
:param prefix: -
:param al_test: if True then eval_dataset is Dpool
:param mc_samples: if not None, int with number of MC forward samples
:return:
"""
# Loop to handle MNLI double evaluation (matched, mis-matched)
eval_task_names = ("mnli", "mnli-mm") if args.task_name == "mnli" else (args.task_name,)
eval_outputs_dirs = (args.output_dir, args.output_dir + "-MM") if args.task_name == "mnli" else (args.output_dir,)
results = {}
for eval_task, eval_output_dir in zip(eval_task_names, eval_outputs_dirs):
# eval_dataset = load_and_cache_examples(args, eval_task, tokenizer, evaluate=True)
if not os.path.exists(eval_output_dir) and args.local_rank in [-1, 0]:
os.makedirs(eval_output_dir)
args.eval_batch_size = args.per_gpu_eval_batch_size * max(1, args.n_gpu)
# Sequential sampler - crucial!!!
eval_sampler = SequentialSampler(eval_dataset)
eval_dataloader = DataLoader(eval_dataset, sampler=eval_sampler, batch_size=args.eval_batch_size)
# multi-gpu eval
if args.n_gpu > 1 and not isinstance(model, torch.nn.DataParallel):
model = torch.nn.DataParallel(model)
# Eval!
logger.info("***** Running evaluation {} *****".format(prefix))
logger.info(" Num examples = %d", len(eval_dataset))
logger.info(" Batch size = %d", args.eval_batch_size)
eval_loss = 0.0
nb_eval_steps = 0
preds = None
out_label_ids = None
bert_output_list = None
if mc_samples is not None:
# MC dropout
test_losses = []
logits_list = []
for i in range(1, mc_samples + 1):
test_losses_mc = []
logits_mc = None
for batch in tqdm(eval_dataloader, desc="Evaluating"):
model.train()
batch = tuple(t.to(args.device) for t in batch)
with torch.no_grad():
inputs = {"input_ids": batch[0], "attention_mask": batch[1], "labels": batch[-1]}
if args.model_type != "distilbert":
inputs["token_type_ids"] = (
batch[2] if args.model_type in ["bert", "xlnet", "albert"] else None
) # XLM, DistilBERT, RoBERTa, and XLM-RoBERTa don't use segment_ids
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()
logits_mc = logits
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)
logits_mc = torch.cat((logits_mc, logits), 0)
test_losses_mc.append(eval_loss / nb_eval_steps)
# logits_mc.append(logits)
test_losses.append(test_losses_mc)
logits_list.append(logits_mc)
preds = None
eval_loss = np.mean(test_losses)
logits = logits_list
preds = torch.mean(torch.stack(logits), 0).detach().cpu().numpy()
else:
# Standard inference (no MC dropout)
for batch in tqdm(eval_dataloader, desc="Evaluating"):
model.eval()
batch = tuple(t.to(args.device) for t in batch)
with torch.no_grad():
inputs = {"input_ids": batch[0], "attention_mask": batch[1], "labels": batch[-1]}
if args.model_type != "distilbert":
inputs["token_type_ids"] = (
batch[2] if args.model_type in ["bert", "xlnet", "albert"] else None
) # XLM, DistilBERT, RoBERTa, and XLM-RoBERTa don't use segment_ids
outputs = model(**inputs)
labels = inputs.pop("labels",None)
if return_bert_embs:
bert_output = model.bert(**inputs)[1]
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()
out_label_ids = labels.detach().cpu().numpy()
if return_bert_embs:
# bert_output_list = bert_output.detach().cpu().numpy()
bert_output_list = bert_output
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)
out_label_ids = np.append(out_label_ids, labels.detach().cpu().numpy(), axis=0)
if return_bert_embs:
# bert_output_list = np.append(bert_output_list, bert_output.detach().cpu().numpy(), axis=0)
bert_output_list = torch.cat((bert_output_list, bert_output),0)
eval_loss = eval_loss / nb_eval_steps
logits = torch.tensor(preds)
if args.output_mode == "classification":
preds = np.argmax(preds, axis=1)
# accuracy = round(acc(out_label_ids, preds), 4)
# f1 = round(f1_macro(out_label_ids, preds), 4)
# precision = round(precision_macro(out_label_ids, preds), 4)
# recall = round(recall_macro(out_label_ids, preds), 4)
# calibration scores
calibration_scores = uncertainty_metrics(logits, out_label_ids,
num_classes=args.num_classes)
elif args.output_mode == "regression":
preds = np.squeeze(preds)
calibration_scores = {}
accuracy, f1, precision, recall = 0., 0., 0., 0.,
calibration_scores = None
result = compute_metrics(eval_task, preds, out_label_ids)
results.update(result)
output_eval_file = os.path.join(eval_output_dir, "eval_results.txt")
with open(output_eval_file, "w") as writer:
logger.info("***** Eval results {} *****".format(prefix))
for key in sorted(result.keys()):
logger.info(" %s = %s", key, str(result[key]))
writer.write("%s = %s\n" % (key, str(result[key])))
# results.update({'f1_macro': f1, 'recall': recall, 'precision': precision, 'loss': eval_loss})
results.update(calibration_scores)
results.update({'bert_output': bert_output_list})
results.update({'gold_labels': out_label_ids.tolist()})
results['loss'] = eval_loss
# return results, eval_loss, accuracy, f1, precision, recall, logits
return results, logits
def train_transformer(args, train_dataset, eval_dataset, model, tokenizer):
"""
Train a transformer model.
:param args: args
:param train_dataset: train (tensor) dataset
:param eval_dataset: dev (tensor) dataset
:param model: model to train
:param tokenizer: tokenizer
:return:
"""
# 10% warmup
args.warmup_steps = int(len(train_dataset) / args.per_gpu_train_batch_size * args.num_train_epochs / 10)
if hasattr(args, "warmup_thr"):
if args.warmup_thr is not None:
args.warmup_steps = min(int(len(train_dataset) / args.per_gpu_train_batch_size * args.num_train_epochs / 10), args.warmup_thr)
print("warmup steps: {}".format(args.warmup_steps))
print("total steps: {}".format(int(len(train_dataset) / args.per_gpu_train_batch_size * args.num_train_epochs)))
print("logging steps: {}".format(args.logging_steps))
##############################
# Train model
##############################
_, model_class, _ = MODEL_CLASSES[args.model_type]
global_step, tr_loss, val_acc, val_loss = train(args, train_dataset, eval_dataset, model, tokenizer)
logger.info(" global_step = %s, average loss = %s", global_step, tr_loss)
# AYTO GIATI TO KANW?
# Evaluation
results = {}
if args.do_eval and args.local_rank in [-1, 0]:
# checkpoints = [args.output_dir]
checkpoints = [args.current_output_dir]
# if args.eval_all_checkpoints:
# checkpoints = list(
# # os.path.dirname(c) for c in sorted(glob.glob(args.output_dir + "/**/" + WEIGHTS_NAME, recursive=True))
# os.path.dirname(c) for c in
# sorted(glob.glob(args.current_output_dir + "/**/" + WEIGHTS_NAME, recursive=True))
# )
# logging.getLogger("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 ""
prefix = checkpoint.split("/")[-1] if checkpoint.find("checkpoint") != -1 else ""
model = model_class.from_pretrained(checkpoint)
model.to(args.device)
result, logits = my_evaluate(eval_dataset, args, model, prefix=prefix)
# result = dict((k + "_{}".format(global_step), v) for k, v in result.items())
# results.update(result)
eval_loss = val_loss
return model, tr_loss, eval_loss, result
if __name__ == '__main__':
import argparse
parser = argparse.ArgumentParser()
##########################################################################
# Setup args
##########################################################################
parser.add_argument("--local_rank",
type=int, default=-1,
help="For distributed training: local_rank")
parser.add_argument("--no_cuda", action="store_true",
help="Avoid using CUDA when available")
parser.add_argument(
"--fp16",
action="store_true",
help="Whether to use 16-bit (mixed) precision (through NVIDIA apex) instead of 32-bit",
)
parser.add_argument(
"--fp16_opt_level",
type=str,
default="O1",
help="For fp16: Apex AMP optimization level selected in ['O0', 'O1', 'O2', and 'O3']."
"See details at https://nvidia.github.io/apex/amp.html",
)
##########################################################################
# Model args
##########################################################################
parser.add_argument("--model_type", default="bert", type=str, help="Pretrained model")
parser.add_argument("--model_name_or_path", default="bert-base-cased", type=str, help="Pretrained ckpt")
parser.add_argument(
"--config_name", default="", type=str, help="Pretrained config name or path if not the same as model_name",
)
parser.add_argument(
"--tokenizer_name",
default="",
type=str,
help="Pretrained tokenizer name or path if not the same as model_name",
)
parser.add_argument(
"--use_fast_tokenizer",
default=True,
type=bool,
help="Whether to use one of the fast tokenizer (backed by the tokenizers library) or not.",
)
parser.add_argument(
"--do_lower_case", action="store_true",
default=False,
help="Set this flag if you are using an uncased model.",
)
# parser.add_argument("--tapt", default=None, type=str,
# help="ckpt of tapt model")
##########################################################################
# Training args
##########################################################################
parser.add_argument("--do_train", default=True, type=bool, help="If true do train")
parser.add_argument("--do_eval", default=True, type=bool, help="If true do evaluation")
parser.add_argument("--overwrite_output_dir", default=True, type=bool, help="If true do evaluation")
parser.add_argument("--per_gpu_train_batch_size", default=4, type=int, help="Batch size per GPU/CPU for training.")
parser.add_argument("--per_gpu_eval_batch_size", default=32, type=int,
help="Batch size per GPU/CPU for evaluation.")
parser.add_argument(
"--gradient_accumulation_steps",
type=int,
default=1,
help="Number of updates steps to accumulate before performing a backward/update pass.",
)
parser.add_argument(
"--num_train_epochs", default=5.0, type=float, help="Total number of training epochs to perform.",
)
parser.add_argument(
"--max_steps",
default=-1,
type=int,
help="If > 0: set total number of training steps to perform. Override num_train_epochs.",
)
parser.add_argument("--warmup_steps", default=0, type=int, help="Linear warmup over warmup_steps.")
parser.add_argument("--logging_steps", type=int, default=50, help="Log every X updates steps.")
parser.add_argument("--save_steps", type=int, default=0, help="Save checkpoint every X updates steps.")
parser.add_argument(
"--eval_all_checkpoints",
action="store_true",
help="Evaluate all checkpoints starting with the same prefix as model_name ending and ending with step number",
)
parser.add_argument("--learning_rate", default=2e-5, type=float, help="The initial learning rate for Adam.")
parser.add_argument("--weight_decay", default=1e-5, type=float, help="Weight decay if we apply some.")
parser.add_argument("--adam_epsilon", default=1e-8, type=float, help="Epsilon for Adam optimizer.")
parser.add_argument("--max_grad_norm", default=1.0, type=float, help="Max gradient norm.")
parser.add_argument("-seed", "--seed", required=False, type=int, help="seed")
parser.add_argument("--indicator", required=False,
default=None,
type=str,
help="experiment indicator")
parser.add_argument("-patience", "--patience", required=False, type=int, help="patience for early stopping (steps)")
parser.add_argument("--use_adapter", required=False, type=bool,
default=False,
help="if True finetune model with added adapter layers")
parser.add_argument("--use_bayes_adapter", required=False, type=bool,
default=False,
help="if True finetune model with added Bayes adapter layers")
parser.add_argument("--unfreeze_adapters", required=False, type=bool,
default=False,
help="if True add adapters and fine-tune all model")
##########################################################################
# Data args
##########################################################################
parser.add_argument("--dataset_name", default=None, required=True, type=str,
help="Dataset [mrpc, ag_news, qnli, sst-2, trec-6]")
parser.add_argument("--data_dir", default=None, required=False, type=str,
help="Datasets folder")
# parser.add_argument("--task_name", default=None, type=str, help="Task [MRPC, AG_NEWS, QNLI, SST-2]")
parser.add_argument("--max_seq_length", default=256, type=int, help="Max sequence length")
##########################################################################
# Uncertainty estimation args
##########################################################################
parser.add_argument("--unc_method",
default="vanilla",
type=str,
help="Choose uncertainty estimation method from "
"[vanilla, mc, ensemble, temp_scale, bayes_adapt, bayes_top]"
)
parser.add_argument("--test_all_uncertainty", required=False, type=bool, default=True,
help=" if True evaluate [vanilla, mc_3, mc_5, mc_10, mc_20, temp_scaling] "
"uncertainty methods for the model")
parser.add_argument("--bayes_output", required=False, type=bool, default=False,
help=" if True add Bayesian classification layer (UA)")
##########################################################################
# Server args
##########################################################################
parser.add_argument("-g", "--gpu", required=False,
default='0', help="gpu on which this experiment runs")
parser.add_argument("-server", "--server", required=True,
default='ford', help="server on which this experiment runs")
args = parser.parse_args()
# Setup
if args.server is 'ford':
os.environ["CUDA_VISIBLE_DEVICES"] = args.gpu
print("\nThis experiment runs on gpu {}...\n".format(args.gpu))
args.n_gpu = 1
args.device = torch.device('cuda:{}'.format(args.gpu))
else:
# Setup CUDA, GPU & distributed training
if args.local_rank == -1 or args.no_cuda:
args.device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
args.n_gpu = 0 if args.no_cuda else 1
else: # Initializes the distributed backend which will take care of sychronizing nodes/GPUs
torch.cuda.set_device(args.local_rank)
args.device = torch.device("cuda", args.local_rank)
torch.distributed.init_process_group(backend="nccl")
args.n_gpu = 1
print('device: {}'.format(args.device))
#########################################
# Setup args
#########################################
if args.seed == None:
seed = random.randint(1, 9999)
args.seed = seed
args.task_name = args.dataset_name.upper()
args.cache_dir = CACHE_DIR
if args.data_dir is None:
args.data_dir = os.path.join(DATA_DIR, args.task_name)
if args.dataset_name == 'cola': args.data_dir = os.path.join(DATA_DIR, "CoLA")
args.overwrite_cache = True
args.evaluate_during_training = True
# Output dir
output_dir = os.path.join(CKPT_DIR, '{}_{}'.format(args.dataset_name, args.model_type))
args.output_dir = os.path.join(output_dir, 'all_{}'.format(args.seed))
if args.use_adapter: args.output_dir += '-adapter'
if args.use_bayes_adapter: args.output_dir += '-bayes-adapter'
if args.indicator is not None: args.output_dir += '-{}'.format(args.indicator)
if args.patience is not None: args.output_dir += '-early{}'.format(int(args.num_train_epochs))
if args.bayes_output: args.output_dir += '-bayes-output'
if (args.use_adapter or args.bayes_output) and args.unfreeze_adapters: args.output_dir += '-unfreeze'
args.current_output_dir = args.output_dir
if (
os.path.exists(args.output_dir)
and os.listdir(args.output_dir)
and args.do_train
and not args.overwrite_output_dir
):
raise ValueError(
"Output directory ({}) already exists and is not empty. Use --overwrite_output_dir to overcome.".format(
args.output_dir
)
)
#########################################
# Setup logging
#########################################
logging.basicConfig(
format="%(asctime)s - %(levelname)s - %(name)s - %(message)s",
datefmt="%m/%d/%Y %H:%M:%S",
level=logging.INFO if args.local_rank in [-1, 0] else logging.WARN,
)
logger.warning(
"Process rank: %s, device: %s, n_gpu: %s, distributed training: %s, 16-bits training: %s",
args.local_rank,
args.device,
args.n_gpu,
bool(args.local_rank != -1),
args.fp16,
)
#########################################
# Prepare GLUE task
#########################################
args.task_name = args.task_name.lower()
if args.task_name not in processors:
raise ValueError("Task not found: %s" % (args.task_name))
processor = processors[args.task_name]()
args.output_mode = output_modes[args.task_name]
label_list = processor.get_labels()
num_labels = len(label_list)
# Load pretrained model and tokenizer
if args.local_rank not in [-1, 0]:
torch.distributed.barrier() # Make sure only the first process in distributed training will download model & vocab
args.model_type = args.model_type.lower()
config_class, model_class, tokenizer_class = MODEL_CLASSES[args.model_type]
config = config_class.from_pretrained(
args.config_name if args.config_name else args.model_name_or_path,
num_labels=num_labels,
finetuning_task=args.task_name,
cache_dir=args.cache_dir if args.cache_dir else None,
use_adapter=args.use_adapter,
use_bayes_adapter=args.use_bayes_adapter,
adapter_initializer_range=0.0002 if args.indicator=='identity_init' else 1,
bayes_output=args.bayes_output,
unfreeze_adapters=args.unfreeze_adapters
)
tokenizer = tokenizer_class.from_pretrained(
args.tokenizer_name if args.tokenizer_name else args.model_name_or_path,
do_lower_case=args.do_lower_case,
cache_dir=args.cache_dir if args.cache_dir else None,
)
model = model_class.from_pretrained(
args.model_name_or_path,
from_tf=bool(".ckpt" in args.model_name_or_path),
config=config,
cache_dir=args.cache_dir if args.cache_dir else None,
)
#########################################
# Check if experiment already done
#########################################
create_dir(RES_DIR)
path = os.path.join(RES_DIR, '{}_{}_100%'.format(args.task_name, args.model_type))
create_dir(path)
name = 'seed_{}_lr_{}_bs_{}_epochs_{}'.format(args.seed, args.learning_rate,
args.per_gpu_train_batch_size,
int(args.num_train_epochs))
# if args.use_adapter: name += '_adapters'
if args.indicator is not None: name += '_{}'.format(args.indicator)
print(name)
dirname = os.path.join(path, name)
if os.path.isdir(dirname) and os.listdir(dirname):
json_file = 'temp_scale_results.json' if not args.bayes_output else 'temp_scale_results_bayes_output.json'
if os.path.isfile(os.path.join(dirname, json_file)):
print('Experiment done!')
exit()
create_dir(dirname)
#########################################
# Load (raw) dataset
#########################################
X_train, y_train = get_glue_dataset(args, args.data_dir, args.task_name, args.model_type, evaluate=False)
X_val, y_val = get_glue_dataset(args, args.data_dir, args.task_name, args.model_type, evaluate=True)
X_test, y_test = get_glue_dataset(args, args.data_dir, args.task_name, args.model_type, test=True)
# test_dataset_ood = get_glue_tensor_dataset(None, args, 'twitterppdb', tokenizer, test=True,
# data_dir=os.path.join(DATA_DIR, 'TwitterPPDB'))
X_orig = X_train # original train set
y_orig = y_train # original labels
X_inds = list(np.arange(len(X_orig))) # indices to original train set
X_unlab_inds = [] # indices of ulabeled set to original train set
args.binary = True if len(set(y_train)) == 2 else False
args.num_classes = len(set(y_train))
# The following code is in case we want to undersample the dataset to evaluate uncertainty
# in low (data) resource scenarios
args.undersampling = False
if args.indicator is not None:
# Undersample training dataset (stratified sampling)
if "sample_" in args.indicator:
args.undersampling = True
num_to_sample = int(args.indicator.split("_")[1])
X_train_orig_after_sampling_inds, X_train_orig_remaining_inds, _, _ = train_test_split(
X_inds,
y_orig,
train_size=num_to_sample,
random_state=args.seed,
stratify=y_train)
X_inds = X_train_orig_after_sampling_inds # indices of train set to original train set
# X_train = list(np.array(X_train)[X_inds]) # train set
# y_train = list(np.array(y_train)[X_inds]) # labels
# Treat the rest of training data as unlabeled data
X_unlab_inds = X_train_orig_remaining_inds # indices of ulabeled set to original train set
assert len(X_unlab_inds) + len(X_inds) == len(X_orig)
assert bool(not (set(X_unlab_inds) & set(X_inds)))
assert max(X_inds) < len(X_orig)
if X_unlab_inds != []:
assert max(X_unlab_inds) < len(X_orig)
#########################################
# Load (tensor) dataset
#########################################
train_dataset = get_glue_tensor_dataset(X_inds, args, args.task_name, tokenizer, train=True)
assert len(train_dataset) == len(X_inds)
eval_dataset = get_glue_tensor_dataset(None, args, args.task_name, tokenizer, evaluate=True)
test_dataset = get_glue_tensor_dataset(None, args, args.task_name, tokenizer, test=True)
if args.dataset_name == 'mnli':
test_dataset_ood = get_glue_tensor_dataset(None, args, args.task_name, tokenizer, test=True, ood=True)
#######################
# Train setup
#######################
# select after how many steps will evaluate during training so that we will evaluate at least 5 times in one epoch
minibatch = int(len(X_inds) / (args.per_gpu_train_batch_size * max(1, args.n_gpu)))
args.logging_steps = min(int(minibatch / 5), 500)
if args.logging_steps < 1:
args.logging_steps = 1
if args.local_rank == 0:
torch.distributed.barrier() # Make sure only the first process in distributed training will download model & vocab
model.to(args.device)
logger.info("Training/evaluation parameters %s", args)
#######################
# Train
#######################
model, tr_loss, val_loss, val_results = train_transformer(args, train_dataset, eval_dataset, model, tokenizer)
#######################
# Test uncertainty
#######################
print('Evaluate uncertainty on dev & test sets....')