-
Notifications
You must be signed in to change notification settings - Fork 11
/
run_al.py
827 lines (718 loc) · 43.4 KB
/
run_al.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
import json
import logging
import math
import os
import pickle
import sys
import time
import numpy as np
import torch
from sklearn import feature_extraction
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.model_selection import train_test_split
from transformers import set_seed, AutoTokenizer, AutoConfig, AutoModelForSequenceClassification
sys.path.append("../../")
sys.path.append(os.path.dirname(os.path.dirname(os.path.dirname(__file__))))
from acquisition.cal import contrastive_acquisition
from acquisition.uncertainty import select_alps, calculate_uncertainty
from utilities.data_loader import get_glue_dataset, get_glue_tensor_dataset
from utilities.preprocessors import output_modes
from utilities.trainers import test_transformer_model, train_transformer_model, my_evaluate
from sys_config import acquisition_functions, CACHE_DIR, DATA_DIR, CKPT_DIR
from utilities.general import create_dir, print_stats, create_exp_dirs
logger = logging.getLogger(__name__)
def al_loop(args):
"""
Main script for the active learning algorithm.
:param args: contains necessary arguments for model, training, data and AL settings
Datasets (lists): X_train_original, y_train_original, X_val, y_val
Indices (lists): X_train_init_inds : inds of first training set (iteration 1)
X_train_current_inds : inds of labeled dataset (iteration i)
X_train_remaining_inds : inds of unlabeled dataset (iteration i)
X_train_original_inds : inds of (full) original training set
"""
#############
# Setup
#############
# Set the random seed manually for reproducibility.
set_seed(args.seed)
##############################################################
# Load data
##############################################################
X_test_ood = None
X_train_original, y_train_original = 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)
if args.task_name == 'imdb':
X_test_ood, y_test_ood = get_glue_dataset(args, os.path.join(DATA_DIR, 'SST-2'), 'sst-2', args.model_type,
test=True)
if args.task_name == 'sst-2':
X_test_ood, y_test_ood = get_glue_dataset(args, os.path.join(DATA_DIR, 'IMDB'), 'imdb', args.model_type,
test=True)
if args.task_name == 'qqp':
X_test_ood, y_test_ood = get_glue_dataset(args, os.path.join(DATA_DIR, 'TwitterPPDB'), 'twitterppdb',
args.model_type, test=True)
X_train_original_inds = list(np.arange(len(X_train_original)))[:args.cap_training_pool] # original pool
X_val_inds = list(np.arange(len(X_val)))
X_test_inds = list(np.arange(len(X_test)))
if args.dataset_name in ['dbpedia']:
# undersample dpool up to 20K + dval up to 2K
new_X_train_original_inds, X_train_discarded_inds, _, _ = train_test_split(X_train_original_inds,
y_train_original,
train_size=20000,
random_state=42,
stratify=y_train_original)
new_X_val_inds, X_val_discarded_inds, _, _ = train_test_split(X_val_inds,
y_val,
train_size=2000,
random_state=42,
stratify=y_val)
X_train_original_inds = new_X_train_original_inds
X_val_inds = new_X_val_inds
args.binary = True if len(set(np.array(y_train_original)[X_train_original_inds])) == 2 else False
args.num_classes = len(set(np.array(y_train_original)[X_train_original_inds]))
# set acquisition_size and init_train_data to their correct values based on if they were meant as a percentage
args.acquisition_size, is_percentage = args.acquisition_size
if is_percentage:
args.acquisition_size = round(len(X_train_original_inds) * args.acquisition_size / 100)
args.init_train_data, is_percentage = args.init_train_data
if is_percentage:
args.init_train_data = round(len(X_train_original_inds) * args.init_train_data / 100)
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,
)
##############################################################
# Stats
##############################################################
print()
print_stats(np.array(y_train_original)[X_train_original_inds], 'train')
print_stats(np.array(y_val)[X_val_inds], 'validation')
print_stats(np.array(y_test)[X_test_inds], 'test')
print(f"\nDataset for annotation: {args.dataset_name}\nAcquisition function: {args.acquisition}"
f"\nBudget: {args.budget[0]}{'% of training set' if args.budget[1] else ' labels'} \n")
init_train_data = args.init_train_data
init_train_percent = init_train_data / len(list(np.array(X_train_original)[X_train_original_inds])) * 100
##############################################################
# Experiment dir
##############################################################
results_per_iteration = {}
results_dir = create_exp_dirs(args)
resume_dir = results_dir
##############################################################
# Get BERT representations
##############################################################
bert_representations = None
if args.bert_rep:
if os.path.isfile(os.path.join(args.data_dir, "bert_representations.pkl")):
print('Load bert representations...')
with open(os.path.join(args.data_dir, "bert_representations.pkl"), 'rb') as handle:
bert_representations = pickle.load(handle)
assert bert_representations.shape[0] == len(X_train_original_inds)
else:
args.task_name = args.task_name.lower()
args.output_mode = output_modes[args.task_name]
ori_dataset = get_glue_tensor_dataset(X_train_original_inds, args, args.task_name, tokenizer, train=True)
bert_config = AutoConfig.from_pretrained(
# args.config_name if args.config_name else args.model_name_or_path,
'bert-base-cased',
num_labels=args.num_classes,
finetuning_task=args.task_name,
cache_dir=args.cache_dir,
)
bert_tokenizer = AutoTokenizer.from_pretrained(
# args.tokenizer_name if args.tokenizer_name else args.model_name_or_path,
'bert-base-cased',
cache_dir=args.cache_dir,
use_fast=args.use_fast_tokenizer,
)
bert_model = AutoModelForSequenceClassification.from_pretrained(
# args.model_name_or_path,
'bert-base-cased',
from_tf=bool(".ckpt" in args.model_name_or_path),
config=bert_config,
cache_dir=args.cache_dir,
)
bert_model.to(args.device)
# eval_loss, logits, result
_, _, _results = test_transformer_model(args, dataset=ori_dataset, model=bert_model, return_cls=True)
bert_representations = _results["bert_cls"]
assert bert_representations.shape[0] == len(X_train_original_inds)
with open(os.path.join(args.data_dir, "bert_representations.pkl"), 'wb') as handle:
pickle.dump(bert_representations, handle, protocol=pickle.HIGHEST_PROTOCOL)
##############################################################
# Get TFIDF representations
##############################################################
tfidf_representations = None
if args.tfidf:
if os.path.isfile(os.path.join(args.data_dir, "tfidf_representations.pkl")):
print('Load tfidf representations...')
with open(os.path.join(args.data_dir, "tfidf_representations.pkl"), 'rb') as handle:
tfidf_representations = pickle.load(handle)
assert len(tfidf_representations) == len(X_train_original_inds)
else:
vectorizer = TfidfVectorizer(max_features=15000, lowercase=True,
stop_words=feature_extraction.text.ENGLISH_STOP_WORDS)
if type(X_train_original[0]) is list or type(X_train_original[0]) is tuple:
vectors = vectorizer.fit_transform(
[s[0] + ' ' + s[1] for s in np.array(X_train_original)[X_train_original_inds]])
else:
vectors = vectorizer.fit_transform([s for s in np.array(X_train_original)[X_train_original_inds]])
feature_names = vectorizer.get_feature_names()
dense = vectors.todense()
denselist = dense.tolist()
tfidf_representations = denselist
# tfidf_representations = torch.tensor(denselist)
assert len(tfidf_representations) == len(X_train_original_inds)
with open(os.path.join(args.data_dir, "tfidf_representations.pkl"), 'wb') as handle:
pickle.dump(tfidf_representations, handle, protocol=pickle.HIGHEST_PROTOCOL)
##############################################################
# Resume
##############################################################
if args.resume:
if not os.path.exists(results_dir) or not os.listdir(results_dir) or len(os.listdir(results_dir)) < 2:
args.resume = False
print('Experiment does not exist. Cannot resume. Start from the beginning.')
if args.resume:
print("Resume AL loop.....")
with open(os.path.join(resume_dir, 'results_of_iteration.json'), 'r') as f:
results_per_iteration = json.load(f)
with open(os.path.join(resume_dir, 'selected_ids_per_iteration.json'), 'r') as f:
ids_per_it = json.load(f)
current_iteration = results_per_iteration['last_iteration'] + 1
X_train_current_inds = []
for key in ids_per_it:
X_train_current_inds += ids_per_it[key]
X_train_remaining_inds = [i for i in X_train_original_inds if i not in X_train_current_inds]
assert len(X_train_current_inds) + len(X_train_remaining_inds) == len(
X_train_original_inds), f"current {len(X_train_current_inds)}, remaining {len(X_train_remaining_inds)}, " \
f"original {len(X_train_original_inds)}"
print(f"Current labeled dataset {len(X_train_current_inds)}")
print(f"Unlabeled dataset (Dpool) {len(X_train_remaining_inds)}")
current_annotations = results_per_iteration['current_annotations']
annotations_per_iteration = results_per_iteration['annotations_per_iteration']
total_annotations, is_percentage = args.budget
if is_percentage:
total_annotations = round(total_annotations * len(X_train_original_inds) / 100)
assert current_annotations <= total_annotations, "Experiment done already!"
total_iterations = round(total_annotations / annotations_per_iteration)
if annotations_per_iteration != args.acquisition_size:
annotations_per_iteration = args.acquisition_size
iterations_left = total_iterations - round(current_annotations / annotations_per_iteration)
print(f"New budget! {iterations_left} more iterations.....")
X_discarded_inds = [x for x in X_train_original_inds if x not in X_train_remaining_inds
and x not in X_train_current_inds]
assert len(X_train_current_inds) + len(X_train_remaining_inds) + len(X_discarded_inds) == \
len(X_train_original_inds), f"current {len(X_train_current_inds)}, " \
f"remaining {len(X_train_remaining_inds)}, " \
f"discarded {len(X_discarded_inds)}, " \
f"original {len(X_train_original_inds)}"
assert bool(not set(X_train_current_inds) & set(X_train_remaining_inds))
it2per = {} # iterations to data percentage
val_acc_previous = None
args.acc_best_iteration = 0
args.acc_best = 0
print(f"current iteration {current_iteration}")
print(f"annotations_per_iteration {annotations_per_iteration}")
print(f"budget {args.budget[0]}{'% of training data' if args.budget[1] else ' labels'}")
else:
##############################################################
# New experiment!
##############################################################
##############################################################
# Denote labeled and unlabeled datasets
##############################################################
# ids_per_iteration dict: contains the indices selected at each AL iteration
ids_per_it = {}
##############################################################
# Select first training data
##############################################################
y_strat = np.array(y_train_original)[X_train_original_inds]
X_train_original_after_sampling_inds = []
X_train_original_after_sampling = []
if args.acquisition == 'alps':
args.init = 'alps'
if args.init == 'random':
X_train_init_inds, X_train_remaining_inds, _, _ = train_test_split(X_train_original_inds,
np.array(y_train_original)[
X_train_original_inds],
# train_size=init_train_percent / 100,
train_size=args.init_train_data,
random_state=args.seed,
stratify=y_strat)
elif args.init == 'alps':
X_train_init_inds = select_alps(args, sampled=[], acquisition_size=args.init_train_data,
original_inds=X_train_original_inds)
X_train_remaining_inds = [x for x in X_train_original_inds if x not in X_train_init_inds]
else:
print(args.init)
raise NotImplementedError
####################################################################
# Create Dpool and Dlabels
####################################################################
X_train_init = list(np.asarray(X_train_original, dtype='object')[X_train_init_inds])
y_train_init = list(np.asarray(y_train_original, dtype='object')[X_train_init_inds])
for i in list(set(y_train_init)):
init_train_dist_class = 100 * np.sum(np.array(y_train_init) == i) / len(y_train_init)
print(f'init % class {i}: {init_train_dist_class}')
if X_train_original_after_sampling_inds == []:
assert len(X_train_init_inds) + len(X_train_remaining_inds) == len(
X_train_original_inds), f'init {len(X_train_init_inds)}, remaining {len(X_train_remaining_inds)}, original {len(X_train_original_inds)}'
else:
assert len(X_train_init_inds) + len(X_train_remaining_inds) == len(X_train_original_after_sampling_inds)
ids_per_it.update({str(0): list(map(int, X_train_init_inds))})
assert len(ids_per_it[str(0)]) == args.init_train_data
####################################################################
# Annotations & budget
####################################################################
current_annotations = len(X_train_init) # without validation data
total_annotations, is_percentage = args.budget
if is_percentage:
if X_train_original_after_sampling == []:
total_annotations = round(
total_annotations * len(np.array(X_train_original)[X_train_original_inds]) / 100)
else:
total_annotations = round(total_annotations * len(X_train_original_after_sampling) / 100)
annotations_per_iteration = args.acquisition_size
total_iterations = math.ceil(total_annotations / annotations_per_iteration)
X_train_current_inds = X_train_init_inds.copy()
X_discarded_inds = [x for x in X_train_original_inds if x not in X_train_remaining_inds
and x not in X_train_current_inds]
it2per = {} # iterations to data percentage
val_acc_previous = None
args.acc_best_iteration = 0
args.acc_best = 0
current_iteration = 1
assert bool(not set(X_train_remaining_inds) & set(X_train_current_inds))
"""
Indices of X_train_original: X_train_init_inds - inds of first training set (iteration 1)
X_train_current_inds - inds of labeled dataset (iteration i)
X_train_remaining_inds - inds of unlabeled dataset (iteration i)
X_train_original_inds - inds of (full) original training set
X_disgarded_inds - inds from dpool that are disgarded
"""
#############
# Start AL!
#############
while current_iteration < total_iterations + 1:
it2per[str(current_iteration)] = round(len(X_train_current_inds) / len(X_train_original_inds), 2) * 100
##############################################################
# Train model on training dataset (Dtrain)
##############################################################
print(f"\n Start Training model of iteration {current_iteration}!\n")
train_results = train_transformer_model(args, X_train_current_inds,
X_val_inds,
iteration=current_iteration,
val_acc_previous=val_acc_previous,
)
val_acc_previous = train_results['acc']
print("\nDone Training!\n")
##############################################################
# Test model on test data (D_test)
##############################################################
print("\nStart Testing on test set!\n")
test_dataset = get_glue_tensor_dataset(None, args, args.task_name, tokenizer, test=True)
test_results, test_logits = my_evaluate(test_dataset, args, train_results['model'], prefix="",
al_test=False, mc_samples=None)
test_results.pop('gold_labels', None)
##############################################################
# Test model on OOD test data (D_ood)
##############################################################
print("\nEvaluating robustness! Start testing on OOD test set!\n")
# if False:
if X_test_ood is not None:
if args.dataset_name == 'sst-2':
ood_test_dataset = get_glue_tensor_dataset(None, args, 'imdb', tokenizer, test=True,
data_dir=os.path.join(DATA_DIR, 'IMDB'))
elif args.dataset_name == 'imdb':
ood_test_dataset = get_glue_tensor_dataset(None, args, 'sst-2', tokenizer, test=True,
data_dir=os.path.join(DATA_DIR, 'SST-2'))
elif args.dataset_name == 'qqp':
ood_test_dataset = get_glue_tensor_dataset(None, args, 'twitterppdb', tokenizer, test=True,
data_dir=os.path.join(DATA_DIR, 'TwitterPPDB'))
else:
ood_test_dataset = get_glue_tensor_dataset(None, args, args.task_name, tokenizer, test=True,
ood=True)
ood_test_results, ood_test_logits = my_evaluate(ood_test_dataset, args, train_results['model'],
prefix="",
al_test=False, mc_samples=None)
ood_test_results.pop('gold_labels', None)
##############################################################
# Test model on unlabeled data (Dpool)
##############################################################
print("\nEvaluating Dpool!\n")
start = time.time()
dpool_loss, logits_dpool, results_dpool = [], [], []
if args.acquisition not in ['random', 'alps', 'badge', 'FTbertKM']:
dpool_loss, logits_dpool, results_dpool = test_transformer_model(args, X_train_remaining_inds,
model=train_results['model'],
return_mean_embs=args.mean_embs,
return_mean_output=args.mean_out,
return_cls=args.cls)
results_dpool.pop('gold_labels', None)
end = time.time()
inference_time = end - start
##############################################################
# Select unlabeled samples for annotation
# -> annotate
# -> update training dataset & unlabeled dataset
##############################################################
# I moved this in the end!
# if total_annotations - current_annotations < annotations_per_iteration:
# annotations_per_iteration = total_annotations - current_annotations
#
# if annotations_per_iteration == 0:
# break
assert len(set(X_train_current_inds)) == len(X_train_current_inds)
assert len(set(X_train_remaining_inds)) == len(X_train_remaining_inds)
start = time.time()
if args.acquisition in ["cal", "contrastive"]:
if args.tfidf:
tfidf_dtrain_reprs = torch.tensor(list(np.array(tfidf_representations)[X_train_current_inds]))
tfidf_dpool_reprs = torch.tensor(list(np.array(tfidf_representations)[X_train_remaining_inds]))
else:
tfidf_dtrain_reprs = None
tfidf_dpool_reprs = None
sampled_ind, stats = contrastive_acquisition(args=args,
annotations_per_iteration=annotations_per_iteration,
X_original=X_train_original,
y_original=y_train_original,
labeled_inds=X_train_current_inds,
candidate_inds=X_train_remaining_inds,
discarded_inds=X_discarded_inds,
original_inds=X_train_original_inds,
tokenizer=tokenizer,
train_results=train_results,
results_dpool=results_dpool,
logits_dpool=logits_dpool,
bert_representations=bert_representations,
tfidf_dtrain_reprs=tfidf_dtrain_reprs,
tfidf_dpool_reprs=tfidf_dpool_reprs)
else:
sampled_ind, stats = calculate_uncertainty(args=args,
method=args.acquisition,
logits=logits_dpool,
annotations_per_it=annotations_per_iteration,
device=args.device,
iteration=current_iteration,
task=args.task_name,
representations=None,
candidate_inds=X_train_remaining_inds,
labeled_inds=X_train_current_inds,
discarded_inds=X_discarded_inds,
original_inds=X_train_original_inds,
model=train_results['model'],
X_original=X_train_original,
y_original=y_train_original)
end = time.time()
selection_time = end - start
# Update results dict
results_per_iteration[str(current_iteration)] = {'data_percent': it2per[str(current_iteration)],
'total_train_samples': len(X_train_current_inds),
'inference_time': inference_time,
'selection_time': selection_time}
results_per_iteration[str(current_iteration)]['val_results'] = train_results
results_per_iteration[str(current_iteration)]['test_results'] = test_results
if X_test_ood is not None:
results_per_iteration[str(current_iteration)]['ood_test_results'] = ood_test_results
results_per_iteration[str(current_iteration)]['ood_test_results'].pop('model', None)
results_per_iteration[str(current_iteration)]['val_results'].pop('model', None)
results_per_iteration[str(current_iteration)]['test_results'].pop('model', None)
results_per_iteration[str(current_iteration)].update(stats)
current_annotations += annotations_per_iteration
# X_train_current_inds and X_train_remaining_inds are lists of indices of the original dataset
# sampled_inds is a list of indices OF THE X_train_remaining_inds(!!!!) LIST THAT SHOULD BE REMOVED
# INCEPTION %&#!@***CAUTION***%&#!@
if args.acquisition in ['alps', 'badge', 'adv', 'FTbertKM', 'adv_train', "cal", "contrastive"]:
X_train_current_inds += list(sampled_ind)
else:
X_train_current_inds += list(np.array(X_train_remaining_inds)[sampled_ind])
assert len(ids_per_it[str(0)]) == args.init_train_data
if args.acquisition in ['alps', 'badge', 'adv', 'FTbertKM', 'adv_train', "cal", "contrastive"]:
selected_dataset_ids = sampled_ind
selected_dataset_ids = list(map(int, selected_dataset_ids)) # for json
assert len(ids_per_it[str(0)]) == args.init_train_data
else:
selected_dataset_ids = list(np.array(X_train_remaining_inds)[sampled_ind])
selected_dataset_ids = list(map(int, selected_dataset_ids)) # for json
assert len(ids_per_it[str(0)]) == args.init_train_data
ids_per_it.update({str(current_iteration): selected_dataset_ids})
assert len(ids_per_it[str(0)]) == args.init_train_data
assert len(ids_per_it[str(current_iteration)]) == annotations_per_iteration
if args.acquisition in ['alps', 'badge', 'adv', 'FTbertKM', 'adv_train', "cal", "contrastive"]:
X_train_remaining_inds = [x for x in X_train_original_inds if x not in X_train_current_inds
and x not in X_discarded_inds]
else:
X_train_remaining_inds = list(np.delete(X_train_remaining_inds, sampled_ind))
# Assert no common data in Dlab and Dpool
assert bool(not set(X_train_current_inds) & set(X_train_remaining_inds))
# Assert unique (no duplicate) inds in Dlab & Dpool
assert len(set(X_train_current_inds)) == len(X_train_current_inds)
assert len(set(X_train_remaining_inds)) == len(X_train_remaining_inds)
# Assert each list of inds unique
set(X_train_original_inds).difference(set(X_train_current_inds))
if args.indicator is None and args.indicator != "small_config":
assert set(X_train_original_inds).difference(set(X_train_current_inds)) == set(
X_train_remaining_inds + X_discarded_inds)
results_per_iteration['last_iteration'] = current_iteration
results_per_iteration['current_annotations'] = current_annotations
results_per_iteration['annotations_per_iteration'] = annotations_per_iteration
results_per_iteration['X_val_inds'] = list(map(int, X_val_inds))
print("\n")
print("*" * 12)
print(f"End of iteration {current_iteration}:")
if 'loss' in test_results.keys():
print(
f"Train loss {train_results['train_loss']}, Val loss {train_results['loss']}, Test loss {test_results['loss']}")
print(f"Annotated {annotations_per_iteration} samples")
print(f"Current labeled (training) data: {len(X_train_current_inds)} samples")
print(f"Remaining budget: {total_annotations - current_annotations} (in samples)")
print("*" * 12)
print()
current_iteration += 1
print('Saving json with the results....')
with open(os.path.join(results_dir, 'results_of_iteration.json'), 'w') as f:
json.dump(results_per_iteration, f)
with open(os.path.join(results_dir, 'selected_ids_per_iteration.json'), 'w') as f:
json.dump(ids_per_it, f)
# Check budget
if total_annotations - current_annotations < annotations_per_iteration:
annotations_per_iteration = total_annotations - current_annotations
if annotations_per_iteration == 0:
break
print('The end!....')
return
class Percentable(object):
"""
Represents a number that can either be a normal float or a percentage
Takes X or X% and returns a tuple with (X, bool is_percentage)"""
def __new__(self, percentable_string):
is_percentage = False
if percentable_string.endswith('%'):
is_percentage = True
percentable_string = percentable_string[:-1]
try:
percentable_string = float(percentable_string)
except ValueError:
print(f"{percentable_string} is not a valid float.")
return percentable_string, is_percentage
if __name__ == '__main__':
import argparse
import random
##########################################################################
# Setup args
##########################################################################
parser = argparse.ArgumentParser()
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.",
)
##########################################################################
# 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=16, type=int, help="Batch size per GPU/CPU for training.")
parser.add_argument("--per_gpu_eval_batch_size", default=256, 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=3, 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("--warmup_thr", default=None, type=int, help="apply min threshold to 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=0.0, 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("-patience", "--patience", required=False, type=int, default=None,
help="patience for early stopping (steps)")
##########################################################################
# Data args
##########################################################################
parser.add_argument("--dataset_name", default=None, required=True, type=str,
help="Dataset [mrpc, ag_news, qnli, sst-2]")
parser.add_argument("--max_seq_length", default=256, type=int, help="Max sequence length")
##########################################################################
# AL args
##########################################################################
parser.add_argument("-acquisition", "--acquisition", required=True,
type=str,
choices=acquisition_functions,
help="Choose an acquisition function to be used for AL.")
parser.add_argument("-budget", "--budget", required=False,
default="15%", type=Percentable,
help="budget as 'X%%' percent of training data, or 'X' without %% for total annotations; 15%% by default")
parser.add_argument("-cap_training_pool", "--cap_training_pool", required=False, default=None, type=int,
help="limits the number of samples in the training pool to the first X entries")
parser.add_argument("-mc_samples", "--mc_samples", required=False, default=None, type=int,
help="number of MC forward passes in calculating uncertainty estimates")
parser.add_argument("--resume", required=False,
default=False,
type=bool,
help="if True resume experiment")
parser.add_argument("--acquisition_size", required=False,
default="2%",
type=Percentable,
help="acquisition size at each AL iteration (as absolute 'X' or percentage of train 'X%%'); if None we sample 2%%")
parser.add_argument("--init_train_data", required=False,
default="1%",
type=Percentable,
help="initial training data for AL (as absolute 'X' or percentage of train 'X%%'); if None we sample 1%%")
parser.add_argument("--indicator", required=False,
default=None,
type=str,
help="Experiment indicator")
parser.add_argument("--init", required=False,
default="random",
type=str,
help="random or alps")
parser.add_argument("--reverse", default=False, type=bool, help="if True choose opposite data points")
##########################################################################
# Contrastive acquisition args
##########################################################################
parser.add_argument("--mean_embs", default=False, type=bool, help="if True use bert mean embeddings for kNN")
parser.add_argument("--mean_out", default=False, type=bool, help="if True use bert mean outputs for kNN")
parser.add_argument("--cls", default=True, type=bool, help="if True use cls embedding for kNN")
# parser.add_argument("--kl_div", default=True, type=bool, help="if True choose KL divergence for scoring")
parser.add_argument("--ce", default=False, type=bool, help="if True choose cross entropy for scoring")
parser.add_argument("--operator", default="mean", type=str, help="operator to combine scores of neighbours")
parser.add_argument("--num_nei", default=10, type=float, help="number of kNN to find")
parser.add_argument("--conf_mask", default=False, type=bool, help="if True mask neighbours with confidence score")
parser.add_argument("--conf_thresh", default=0., type=float, help="confidence threshold")
parser.add_argument("--knn_lab", default=False, type=bool, help="if True queries are unlabeled data"
"else labeled")
parser.add_argument("--bert_score", default=False, type=bool, help="if True use bertscore similarity")
parser.add_argument("--tfidf", default=False, type=bool, help="if True use tfidf scores")
parser.add_argument("--bert_rep", default=False, type=bool, help="if True use bert embs (pretrained) similarity")
##########################################################################
# Server args
##########################################################################
parser.add_argument("-g", "--gpu", required=False, default='0', help="gpu on which this experiment runs")
parser.add_argument("-server", "--server", required=False, default='ford',
help="server on which this experiment runs")
# parser.add_argument("--debug", required=False, default=False, help="debug mode")
args = parser.parse_args()
# Setup CUDA, GPU & distributed training
if args.local_rank == -1 or args.no_cuda:
args.device = torch.device("cuda:0" 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(f'device: {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
args.data_dir = os.path.join(DATA_DIR, args.task_name)
args.overwrite_cache = bool(True)
args.evaluate_during_training = True
# Output dir
ckpt_dir = os.path.join(CKPT_DIR,
f'{args.dataset_name}_{args.model_type}_{args.acquisition}_{args.seed}')
args.output_dir = os.path.join(ckpt_dir, f'{args.dataset_name}_{args.model_type}')
if args.model_type == 'allenai/scibert': args.output_dir = os.path.join(ckpt_dir,
f'{args.dataset_name}_{"bert"}') # TODO verify that 'bert' is correct here
if args.indicator is not None: args.output_dir += f'-{args.indicator}'
# The following arguments are experiments in the ablation/analysis section of the paper
if args.reverse: args.output_dir += '-reverse'
if args.mean_embs: args.output_dir += '-inputs'
if args.mean_out: args.output_dir += '-outputs'
if args.cls: args.output_dir += '-cls'
if args.ce: args.output_dir += '-ce'
if args.operator != "mean" and args.acquisition == "adv_train": args.output_dir += f'-{args.operator}'
if args.knn_lab: args.output_dir += '-lab'
if args.bert_score: args.output_dir += '-bs'
if args.bert_rep: args.output_dir += '-br'
if args.tfidf: args.output_dir += '-tfidf'
print(f'output_dir={args.output_dir}')
create_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(
f"Output directory ({args.output_dir}) already exists and is not empty. Use --overwrite_output_dir to overcome."
)
# 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,
)
args.task_name = args.task_name.lower()
al_loop(args)