-
Notifications
You must be signed in to change notification settings - Fork 5
/
train_semi.py
763 lines (663 loc) · 30.9 KB
/
train_semi.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
import argparse
import yaml
import os, sys
import os.path as osp
import pprint
import time
import pickle
import random
import numpy as np
import pandas as pd
import torch
import torch.backends.cudnn as cudnn
import torch.distributed as dist
import torch.nn.functional as F
from torch.utils.tensorboard import SummaryWriter
from imas.dataset.cutmix_tensor import cut_mix, cut_mix_using_v, cut_mix_by_hardness_for_prob, cut_mix_by_hardness, cut_mix_by_hardness_beta
from imas.dataset.builder import get_loader
from imas.models.model_helper import ModelBuilder
from imas.utils.dist_helper import setup_distributed
from imas.utils.loss_helper import get_criterion, compute_unsupervised_loss_by_threshold, compute_ulb_hardness_all
from imas.utils.lr_helper import get_optimizer, get_scheduler
from imas.utils.utils import AverageMeter, intersectionAndUnion, load_state, label_onehot
from imas.utils.utils import init_log, get_rank, get_world_size, set_random_seed, setup_default_logging
from imas.dataset.hardness import HardnessWrite
import warnings
warnings.filterwarnings('ignore')
def main(in_args):
args = in_args
if args.seed is not None:
# print("set random seed to", args.seed)
set_random_seed(args.seed, deterministic=True)
# set_random_seed(args.seed)
cfg = yaml.load(open(args.config, "r"), Loader=yaml.Loader)
# cudnn.enabled = True
# cudnn.benchmark = True
rank, word_size = setup_distributed(port=args.port)
###########################
# 1. output settings
###########################
cfg["exp_path"] = osp.dirname(args.config)
cfg["save_path"] = osp.join(cfg["exp_path"], cfg["saver"]["snapshot_dir"])
cfg["log_path"] = osp.join(cfg["exp_path"], "log")
flag_use_tb = cfg["saver"]["use_tb"]
if not os.path.exists(cfg["log_path"]) and rank == 0:
os.makedirs(cfg["log_path"])
if not osp.exists(cfg["save_path"]) and rank == 0:
os.makedirs(cfg["save_path"])
# setup logger and csv paths
if rank == 0:
logger, curr_timestr = setup_default_logging("global", cfg["log_path"])
csv_path = os.path.join(cfg["log_path"], "seg_{}_stat.csv".format(curr_timestr))
csv_path_for_hardness = os.path.join(cfg["log_path"], "seg_{}_hardness.csv".format(curr_timestr))
else:
logger, curr_timestr = None, ""
csv_path, csv_path_for_hardness = None, None
# make sure all folders and csv handler are correctly created on rank ==0.
dist.barrier()
# create hardness instance
if "pascal" in cfg["dataset"]["type"]:
num_ulb = 10582 - cfg["dataset"]["n_sup"]
elif "cityscapes" in cfg["dataset"]["type"]:
num_ulb = 2975 - cfg["dataset"]["n_sup"]
else:
ValueError
if rank == 0:
logger.info("{}".format(pprint.pformat(cfg)))
if flag_use_tb:
tb_logger = SummaryWriter(
osp.join(cfg["log_path"], "events_seg",curr_timestr)
)
else:
tb_logger = None
else:
tb_logger = None
###########################
# 2. prepare model 1
###########################
model = ModelBuilder(cfg["net"])
modules_back = [model.encoder]
if cfg["net"].get("aux_loss", False):
modules_head = [model.auxor, model.decoder]
else:
modules_head = [model.decoder]
if cfg["net"].get("sync_bn", True):
model = torch.nn.SyncBatchNorm.convert_sync_batchnorm(model)
model.cuda()
###########################
# 3. data
###########################
sup_loss_fn = get_criterion(cfg)
train_loader_sup, train_loader_unsup, val_loader = get_loader(cfg, seed=args.seed)
##############################
# 4. optimizer & scheduler
##############################
cfg_trainer = cfg["trainer"]
cfg_optim = cfg_trainer["optimizer"]
times = 10 if "pascal" in cfg["dataset"]["type"] else 1
params_list = []
for module in modules_back:
params_list.append(
dict(params=module.parameters(), lr=cfg_optim["kwargs"]["lr"])
)
for module in modules_head:
params_list.append(
dict(params=module.parameters(), lr=cfg_optim["kwargs"]["lr"] * times)
)
optimizer = get_optimizer(params_list, cfg_optim)
###########################
# 5. prepare model more
###########################
local_rank = int(os.environ["LOCAL_RANK"])
model = torch.nn.parallel.DistributedDataParallel(
model,
device_ids=[local_rank],
output_device=local_rank,
find_unused_parameters=False,
)
# Teacher model
model_teacher = ModelBuilder(cfg["net"])
model_teacher.cuda()
model_teacher = torch.nn.parallel.DistributedDataParallel(
model_teacher,
device_ids=[local_rank],
output_device=local_rank,
find_unused_parameters=False,
)
for p in model_teacher.parameters():
p.requires_grad = False
# initialize teacher model -- not neccesary if using warmup
with torch.no_grad():
for t_params, s_params in zip(model_teacher.parameters(), model.parameters()):
t_params.data = s_params.data
######################################
# 6. resume
######################################
last_epoch = 0
best_prec = 0
best_epoch = -1
best_prec_stu = 0
best_epoch_stu = -1
# auto_resume > pretrain
if cfg["saver"].get("auto_resume", False):
lastest_model = os.path.join(cfg["save_path"], "ckpt.pth")
if not os.path.exists(lastest_model):
"No checkpoint found in '{}'".format(lastest_model)
else:
print(f"Resume model from: '{lastest_model}'")
best_prec, last_epoch = load_state(
lastest_model, model, optimizer=optimizer, key="model_state"
)
_, _ = load_state(
lastest_model, model_teacher, optimizer=optimizer, key="teacher_state"
)
optimizer_start = get_optimizer(params_list, cfg_optim)
lr_scheduler = get_scheduler(
cfg_trainer, len(train_loader_sup), optimizer_start, start_epoch=last_epoch
)
######################################
# 7. training loop
######################################
if rank == 0:
logger.info('-------------------------- start training --------------------------')
# Start to train model
for epoch in range(last_epoch, cfg_trainer["epochs"]):
# Training
res_loss_sup, res_loss_unsup, res_hardness_dict = train(
model,
model_teacher,
optimizer,
lr_scheduler,
sup_loss_fn,
train_loader_sup,
train_loader_unsup,
epoch,
tb_logger,
logger,
cfg
)
# Update hardness
if rank == 0 and epoch > cfg["trainer"].get("sup_only_epoch", 0):
# record hardness for further analysis
tmp_lst_hardness = {str(x):res_hardness_dict.get(x, None) for x in range(num_ulb)}
tmp_df_hardness = pd.DataFrame(data=tmp_lst_hardness, index=range(epoch, epoch+1))
if epoch > 0 and osp.exists(csv_path_for_hardness):
tmp_df_hardness.to_csv(csv_path_for_hardness, mode='a', header=None, index_label='epoch')
else:
tmp_df_hardness.to_csv(csv_path_for_hardness, index_label='epoch')
# # make sure hardness is updated!!
# dist.barrier()
# Validation
# prec_stu = validate(model, val_loader, epoch, logger, cfg)
# prec_tea = validate(model_teacher, val_loader, epoch, logger, cfg)
prec_stu = validate_citys(model, val_loader, epoch, logger, cfg)
prec_tea = validate_citys(model_teacher, val_loader, epoch, logger, cfg)
prec = prec_tea
if rank == 0:
state = {
"epoch": epoch + 1,
"model_state": model.state_dict(),
"optimizer_state": optimizer.state_dict(),
"teacher_state": model_teacher.state_dict(),
"best_miou": best_prec,
}
if prec_stu > best_prec_stu:
best_prec_stu = prec_stu
best_epoch_stu = epoch
if prec > best_prec:
best_prec = prec
best_epoch = epoch
state["best_miou"] = prec
torch.save(state, osp.join(cfg["save_path"], "ckpt_best.pth"))
torch.save(state, osp.join(cfg["save_path"], "ckpt.pth"))
# save statistics
tmp_results = {
'loss_lb': res_loss_sup,
'loss_ub': res_loss_unsup,
'miou_stu': prec_stu,
'miou_tea': prec_tea,
"best": best_prec,
"best-stu":best_prec_stu}
data_frame = pd.DataFrame(data=tmp_results, index=range(epoch, epoch+1))
if epoch > 0 and osp.exists(csv_path):
data_frame.to_csv(csv_path, mode='a', header=None, index_label='epoch')
else:
data_frame.to_csv(csv_path, index_label='epoch')
logger.info(" <<Test>> - Epoch: {}. MIoU: {:.2f}/{:.2f}. \033[34mBest-STU:{:.2f}/{} \033[31mBest-EMA: {:.2f}/{}\033[0m".format(epoch,
prec_stu * 100, prec_tea * 100, best_prec_stu * 100, best_epoch_stu, best_prec * 100, best_epoch))
if tb_logger is not None:
tb_logger.add_scalar("mIoU val", prec, epoch)
def train(
model,
model_teacher,
optimizer,
lr_scheduler,
sup_loss_fn,
loader_l,
loader_u,
epoch,
tb_logger,
logger,
cfg,
):
ema_decay_origin = cfg["net"]["ema_decay"]
rank, world_size = dist.get_rank(), dist.get_world_size()
model.train()
# data loader
loader_l.sampler.set_epoch(epoch)
loader_u.sampler.set_epoch(epoch)
loader_l_iter = iter(loader_l)
loader_u_iter = iter(loader_u)
assert len(loader_l) == len(
loader_u
), f"labeled data {len(loader_l)} unlabeled data {len(loader_u)}, imbalance!"
# metric indicators
sup_losses = AverageMeter(20)
uns_losses = AverageMeter(20)
batch_times = AverageMeter(20)
learning_rates = AverageMeter(20)
meter_high_pseudo_ratio = AverageMeter(20)
# print freq 4 times for a epoch
print_freq = len(loader_u) // 8 # 8 for semi 4 for sup
print_freq_lst = [i * print_freq for i in range(1,8)]
print_freq_lst.append(len(loader_u) -1)
# create the hardness return
dict_hardness = dict()
# start iterations
model.train()
model_teacher.eval()
for step in range(len(loader_l)):
batch_start = time.time()
i_iter = epoch * len(loader_l) + step # total iters till now
lr = lr_scheduler.get_lr()
learning_rates.update(lr[0])
lr_scheduler.step() # lr is updated at the iteration level
# obtain labeled and unlabeled data
_, image_l, label_l = loader_l_iter.next()
batch_size, h, w = label_l.size()
image_l, label_l = image_l.cuda(), label_l.cuda()
index_u, image_u_weak, image_u_aug, _ = loader_u_iter.next()
index_u, image_u_weak, image_u_aug = index_u.cuda(), image_u_weak.cuda(), image_u_aug.cuda()
# start the training
if epoch < cfg["trainer"].get("sup_only_epoch", 0):
# forward
pred, aux = model(image_l)
# supervised loss
if "aux_loss" in cfg["net"].keys():
sup_loss = sup_loss_fn([pred, aux], label_l)
del aux
else:
sup_loss = sup_loss_fn(pred, label_l)
del pred
# no unlabeled data during the warmup period
unsup_loss = torch.tensor(0.0).cuda()
tensor_hardness = None
pseduo_high_ratio = torch.tensor(0.0).cuda()
else:
# 1. generate pseudo labels and hardness firstly
p_threshold = cfg["trainer"]["unsupervised"].get("threshold", 0.95)
with torch.no_grad():
model.eval()
pred_u_stu, _ = model(image_u_weak.detach())
model_teacher.eval()
pred_u, _ = model_teacher(image_u_weak.detach())
# # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # #
if cfg["dataset"]["train"].get("hardness_aware", False):
flag_using_hardnness = True
# setup detials of using hardness
flag_cal_iou_weighted = cfg["dataset"]["train"]["hardness_aware"].get("flag_cal_iou_weighted", True)
flag_cal_iou_ignore_bg = cfg["dataset"]["train"]["hardness_aware"].get("flag_cal_iou_ignore_bg", True)
flag_using_v1 = cfg["dataset"]["train"]["hardness_aware"].get("flag_using_v1", True)
flag_hardness_weighted_loss = cfg["dataset"]["train"]["hardness_aware"].get("flag_hardness_weighted_loss", True)
flag_cmix_trigger_by_hard = cfg["dataset"]["train"]["hardness_aware"].get("flag_cmix_trigger_by_hard", False)
flag_augs_mixup_by_hard = cfg["dataset"]["train"]["hardness_aware"].get("flag_augs_mixup_by_hard", False)
flag_mapping_random = cfg["dataset"]["train"]["hardness_aware"].get("flag_mapping_random", False)
flag_mapping_gaussian = cfg["dataset"]["train"]["hardness_aware"].get("flag_mapping_gaussian", False)
else:
flag_using_hardnness = False
# still record hardness even if not using hardness
flag_cal_iou_weighted = True
flag_cal_iou_ignore_bg = True
flag_using_v1 = True
flag_hardness_weighted_loss = False
flag_cmix_trigger_by_hard = False
flag_augs_mixup_by_hard = False
flag_mapping_gaussian = False
flag_mapping_random = False
hardness_v1, hardness_v2, hardness_ratio = compute_ulb_hardness_all(pred_u_stu, pred_u, p_threshold,
flag_using_cls_weighted_iou=flag_cal_iou_weighted,
flag_ignoring_background=flag_cal_iou_ignore_bg)
pseduo_high_ratio = hardness_ratio.mean()
if cfg["trainer"]["unsupervised"].get("flag_ema_pseudo", True):
pred_u = F.softmax(pred_u, dim=1)
logits_u_aug, label_u_aug = torch.max(pred_u, dim=1)
del pred_u, pred_u_stu
else:
pred_u_stu = F.softmax(pred_u_stu, dim=1)
logits_u_aug, label_u_aug = torch.max(pred_u_stu, dim=1)
del pred_u, pred_u_stu
# pred_u = F.softmax(pred_u, dim=1)
# pred_u_stu = F.softmax(pred_u_stu, dim=1)
# # pred_u_mix = 0.5 * pred_u + 0.5 * pred_u_stu
# pred_u_mix = ema_decay_origin * pred_u + (1 - ema_decay_origin) * pred_u_stu
# logits_u_aug, label_u_aug = torch.max(pred_u_mix, dim=1)
# del pred_u, pred_u_stu, pred_u_mix
model.train()
# # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # #
# 2. obtain curr and hist hardness
if flag_using_v1:
tensor_hardness = hardness_v1
else:
tensor_hardness = hardness_v2
cur_hardness = tensor_hardness.detach().cpu().numpy().tolist()
hardness_avg = np.mean(cur_hardness)
# 3. apply mixup between strong and weak
flag_loss_intensity_aug = False
if cfg["dataset"]["train"].get("strong_aug", False):
flag_loss_intensity_aug = True
if flag_using_hardnness:
if flag_augs_mixup_by_hard:
mixup_ratio = np.array(cur_hardness)
mixup_ratio = mixup_ratio.reshape((-1,1,1,1)).astype(np.float32)
mixup_ratio = torch.from_numpy(mixup_ratio).cuda()
# image_u_weak ==> cur, image_u_aug ===> past
image_u_aug = image_u_aug * mixup_ratio + image_u_weak * (1.0 - mixup_ratio)
else:
cur_hardness_arr = np.array(cur_hardness)
generate_radomness = np.random.uniform(size=cur_hardness_arr.shape)
mask = generate_radomness < cur_hardness_arr
mask_float = mask.reshape((-1,1,1,1)).astype(np.float32)
mask_float = torch.from_numpy(mask_float).cuda()
image_u_aug = image_u_aug * mask_float + image_u_weak * (1.0 - mask_float)
# 4. apply cutmix on image_u_weak
flag_loss_cutmix_aug = False
if flag_cmix_trigger_by_hard:
trigger_prob = hardness_avg
else:
trigger_prob = cfg["trainer"]["unsupervised"].get("use_cutmix_trigger_prob", 1.0)
if np.random.uniform(0, 1) < trigger_prob and cfg["trainer"]["unsupervised"].get("use_cutmix", False):
flag_loss_cutmix_aug = True
if cfg["trainer"]["unsupervised"].get("use_cutmix_beta", False):
image_u_aug_cm, label_u_aug_cm, logits_u_aug_cm = cut_mix(
image_u_weak,
label_u_aug,
logits_u_aug,
)
else:
c_range = cfg["trainer"]["unsupervised"].get("use_cutmix_range", [0.3, 1/3])
if flag_using_hardnness:
image_u_aug_cm, label_u_aug_cm, logits_u_aug_cm = cut_mix_by_hardness(
image_u_weak,
label_u_aug,
logits_u_aug,
hardness=cur_hardness,
scale=c_range,
flag_hardness_random=flag_mapping_random,
flag_hardness_gaussion=flag_mapping_gaussian)
else:
image_u_aug_cm, label_u_aug_cm, logits_u_aug_cm = cut_mix_using_v(
image_u_weak,
label_u_aug,
logits_u_aug,
scale=c_range
)
# 5. forward concated labeled + unlabeld into student networks
num_labeled = len(image_l)
if flag_loss_cutmix_aug:
if flag_loss_intensity_aug:
pred_all, aux_all = model(torch.cat((image_l, image_u_weak, image_u_aug_cm, image_u_aug), dim=0))
del image_l, image_u_weak, image_u_aug_cm, image_u_aug
pred_l= pred_all[:num_labeled]
pred_u_weak, pred_u_strong_cm, pred_u_strong = pred_all[num_labeled:].chunk(3)
del pred_all
else:
pred_all, aux_all = model(torch.cat((image_l, image_u_weak, image_u_aug_cm), dim=0))
del image_l, image_u_weak, image_u_aug_cm
pred_l= pred_all[:num_labeled]
pred_u_weak, pred_u_strong_cm = pred_all[num_labeled:].chunk(2)
del pred_all
else:
pred_all, aux_all = model(torch.cat((image_l, image_u_weak, image_u_aug), dim=0))
del image_l, image_u_weak, image_u_aug
pred_l= pred_all[:num_labeled]
pred_u_weak, pred_u_strong = pred_all[num_labeled:].chunk(2)
del pred_all
# 6. supervised loss
if "aux_loss" in cfg["net"].keys():
aux = aux_all[:num_labeled]
sup_loss = sup_loss_fn([pred_l, aux], label_l)
del aux_all, aux
else:
# sup_loss = sup_loss_fn(pred_l, label_l.clone())
sup_loss = sup_loss_fn(pred_l, label_l)
# 7. unsupervised loss
if flag_hardness_weighted_loss:
input_hardness = tensor_hardness
else:
input_hardness = None
if flag_loss_cutmix_aug and flag_loss_intensity_aug:
unsup_loss_aug = compute_unsupervised_loss_by_threshold(
pred_u_strong, label_u_aug.detach(),
logits_u_aug.detach(), thresh=p_threshold, hardness_tensor=input_hardness)
unsup_loss_cm = compute_unsupervised_loss_by_threshold(
pred_u_strong_cm, label_u_aug_cm.detach(),
logits_u_aug_cm.detach(), thresh=p_threshold, hardness_tensor=input_hardness)
unsup_loss = (unsup_loss_aug + unsup_loss_cm) * 0.5
del pred_l, pred_u_strong, pred_u_weak, label_u_aug, logits_u_aug, label_u_aug_cm, logits_u_aug_cm
elif flag_loss_cutmix_aug:
unsup_loss = compute_unsupervised_loss_by_threshold(
pred_u_strong_cm, label_u_aug_cm.detach(),
logits_u_aug_cm.detach(), thresh=p_threshold, hardness_tensor=input_hardness)
del pred_l, pred_u_weak, label_u_aug, logits_u_aug, label_u_aug_cm, logits_u_aug_cm
elif flag_loss_intensity_aug:
unsup_loss = compute_unsupervised_loss_by_threshold(
pred_u_strong, label_u_aug.detach(),
logits_u_aug.detach(), thresh=p_threshold, hardness_tensor=input_hardness)
del pred_l, pred_u_strong, pred_u_weak, label_u_aug, logits_u_aug
else:
unsup_loss = compute_unsupervised_loss_by_threshold(
pred_u_strong, label_u_aug.detach(),
logits_u_aug.detach(), thresh=p_threshold, hardness_tensor=input_hardness)
del pred_l, pred_u_strong, pred_u_weak, label_u_aug, logits_u_aug
unsup_loss *= cfg["trainer"]["unsupervised"].get("loss_weight", 1.0)
loss = sup_loss + unsup_loss
# 8. update student model
optimizer.zero_grad()
loss.backward()
optimizer.step()
# 9. update teacher model with EMA
with torch.no_grad():
if epoch > cfg["trainer"].get("sup_only_epoch", 0):
ema_decay = min(
1
- 1
/ (
i_iter
- len(loader_l) * cfg["trainer"].get("sup_only_epoch", 0)
+ 1
),
ema_decay_origin,
)
else:
ema_decay = 0.0
# print("="*20, i_iter, ema_decay)
# update bn of teachers
for param_train, param_eval in zip(model.parameters(), model_teacher.parameters()):
# param_eval.copy_(param_eval * ema_decay + param_train.detach() * (1 - ema_decay))
param_eval.data = param_eval.data * ema_decay + param_train.data * (1 - ema_decay)
for buffer_train, buffer_eval in zip(model.buffers(), model_teacher.buffers()):
# buffer_eval.copy_(buffer_train)
buffer_eval.data = buffer_eval.data * ema_decay + buffer_train.data * (1 - ema_decay)
# buffer_eval.data = buffer_train.data
# 10. gather all loss from different gpus
reduced_sup_loss = sup_loss.clone().detach()
dist.all_reduce(reduced_sup_loss)
sup_losses.update(reduced_sup_loss.item() / world_size)
reduced_uns_loss = unsup_loss.clone().detach()
dist.all_reduce(reduced_uns_loss)
uns_losses.update(reduced_uns_loss.item() / world_size)
reduced_pseudo_high_ratio = pseduo_high_ratio.clone().detach()
dist.all_reduce(reduced_pseudo_high_ratio)
meter_high_pseudo_ratio.update(reduced_pseudo_high_ratio.item() / world_size)
# 11. gather all hardness from different gpus
log_hardness_avg = 0.0
if tensor_hardness is not None:
index_u_all = [torch.zeros_like(index_u) for _ in range(world_size)]
dist.all_gather(index_u_all, index_u)
hardness_all = [torch.zeros_like(tensor_hardness) for _ in range(world_size)]
dist.all_gather(hardness_all, tensor_hardness)
if rank == 0:
# print("="*50)
# print(len(index_u_all), index_u_all[0].shape)
# print(index_u_all)
index_u_all = torch.cat(index_u_all)
hardness_all = torch.cat(hardness_all)
log_hardness_avg = hardness_all.mean().item()
# print(index_u_all.shape, hardness_all.shape)
# print(index_u_all)
# print(hardness_all)
# print("="*50)
index_u_all = index_u_all.cpu().numpy().tolist()
hardness_all = hardness_all.cpu().numpy().tolist()
tmp_dict = {x:y for x,y in zip(index_u_all, hardness_all)}
dict_hardness.update(tmp_dict)
# 12. print log information
batch_end = time.time()
batch_times.update(batch_end - batch_start)
# if i_iter % 10 == 0 and rank == 0:
if step in print_freq_lst and rank == 0:
logger.info(
"Epoch/Iter [{}:{:3}/{:3}]. "
"AvgHard:{:.3}. "
"Sup:{sup_loss.val:.3f}({sup_loss.avg:.3f}) "
"Uns:{uns_loss.val:.3f}({uns_loss.avg:.3f}) "
"Pseudo:{high_ratio.val:.3f}({high_ratio.avg:.3f}) "
"Time:{batch_time.avg:.2f} "
"LR:{lr.val:.5f}".format(
cfg["trainer"]["epochs"], epoch, step,
log_hardness_avg,
# i_iter, cfg["trainer"]["epochs"] * len(loader_l),
sup_loss=sup_losses,
uns_loss=uns_losses,
high_ratio=meter_high_pseudo_ratio,
batch_time=batch_times,
lr=learning_rates,
)
)
if tb_logger is not None:
tb_logger.add_scalar("lr", learning_rates.avg, i_iter)
tb_logger.add_scalar("Sup Loss", sup_losses.avg, i_iter)
tb_logger.add_scalar("Uns Loss", uns_losses.avg, i_iter)
tb_logger.add_scalar("Pseudo ratio", meter_high_pseudo_ratio.avg, i_iter)
return sup_losses.avg, uns_losses.avg, dict_hardness
def validate(
model,
data_loader,
epoch,
logger,
cfg
):
model.eval()
data_loader.sampler.set_epoch(epoch)
num_classes, ignore_label = (
cfg["net"]["num_classes"],
cfg["dataset"]["ignore_label"],
)
rank, world_size = dist.get_rank(), dist.get_world_size()
intersection_meter = AverageMeter()
union_meter = AverageMeter()
for step, batch in enumerate(data_loader):
_, images, labels = batch
images = images.cuda()
labels = labels.long().cuda()
with torch.no_grad():
output, _ = model(images)
# get the output produced by model_teacher
output = output.data.max(1)[1].cpu().numpy()
target_origin = labels.cpu().numpy()
# start to calculate miou
intersection, union, target = intersectionAndUnion(
output, target_origin, num_classes, ignore_label
)
# gather all validation information
reduced_intersection = torch.from_numpy(intersection).cuda()
reduced_union = torch.from_numpy(union).cuda()
reduced_target = torch.from_numpy(target).cuda()
dist.all_reduce(reduced_intersection)
dist.all_reduce(reduced_union)
dist.all_reduce(reduced_target)
intersection_meter.update(reduced_intersection.cpu().numpy())
union_meter.update(reduced_union.cpu().numpy())
iou_class = intersection_meter.sum / (union_meter.sum + 1e-10)
mIoU = np.mean(iou_class)
if rank == 0:
for i, iou in enumerate(iou_class):
logger.info(" [Test] - class [{}] IoU {:.2f}".format(i, iou * 100))
return mIoU
def validate_citys(
model,
data_loader,
epoch,
logger,
cfg
):
model.eval()
data_loader.sampler.set_epoch(epoch)
num_classes = cfg["net"]["num_classes"]
ignore_label = cfg["dataset"]["ignore_label"]
if cfg["dataset"]["val"].get("crop", False):
crop_size,_ = cfg["dataset"]["val"]["crop"].get("size", [769, 769])
else:
crop_size = 769
rank, world_size = dist.get_rank(), dist.get_world_size()
intersection_meter = AverageMeter()
union_meter = AverageMeter()
for step, batch in enumerate(data_loader):
_, images, labels = batch
images = images.cuda()
labels = labels.long()
batch_size, h, w = labels.shape
with torch.no_grad():
final = torch.zeros(batch_size, num_classes, h, w).cuda()
row = 0
while row < h:
col = 0
while col < w:
pred, _ = model(images[:, :, row: min(h, row + crop_size), col: min(w, col + crop_size)])
final[:, :, row: min(h, row + crop_size), col: min(w, col + crop_size)] += pred.softmax(dim=1)
col += int(crop_size * 2 / 3)
row += int(crop_size * 2 / 3)
# get the output
output = final.argmax(dim=1).cpu().numpy()
target_origin = labels.numpy()
# print("="*50, output.shape, output.dtype, target_origin.shape, target_origin.dtype)
# start to calculate miou
intersection, union, target = intersectionAndUnion(
output, target_origin, num_classes, ignore_label
)
# # return ndarray, b*clas
# print("="*20, type(intersection), type(union), type(target), intersection, union, target)
# gather all validation information
reduced_intersection = torch.from_numpy(intersection).cuda()
reduced_union = torch.from_numpy(union).cuda()
reduced_target = torch.from_numpy(target).cuda()
dist.all_reduce(reduced_intersection)
dist.all_reduce(reduced_union)
dist.all_reduce(reduced_target)
intersection_meter.update(reduced_intersection.cpu().numpy())
union_meter.update(reduced_union.cpu().numpy())
iou_class = intersection_meter.sum / (union_meter.sum + 1e-10)
mIoU = np.mean(iou_class)
if rank == 0:
for i, iou in enumerate(iou_class):
logger.info(" [Test] - class [{}] IoU {:.2f}".format(i, iou * 100))
# logger.info(" - <<Test>> - epoch {} mIoU {:.2f}".format(epoch, mIoU * 100))
return mIoU
if __name__ == "__main__":
parser = argparse.ArgumentParser(description="Semi-Supervised Semantic Segmentation")
parser.add_argument("--config", type=str, default="config.yaml")
parser.add_argument("--local_rank", type=int, default=0)
parser.add_argument("--seed", type=int, default=0)
parser.add_argument("--port", default=None, type=int)
args = parser.parse_args()
main(args)